diff --git a/.gitignore b/.gitignore
index 7499b52..5200c61 100644
--- a/.gitignore
+++ b/.gitignore
@@ -5,3 +5,6 @@ MANIFEST
 *.egg-info
 *.pyc
 *~
+
+# Ignore mprof generated files
+mprofile_*.dat
diff --git a/README.rst b/README.rst
index 1cbfee1..84947a7 100644
--- a/README.rst
+++ b/README.rst
@@ -107,10 +107,10 @@ decorator function.  Use as follows::
         del b
         return a
 
-If a python script with decorator ``@profile`` is called using ``-m 
+If a python script with decorator ``@profile`` is called using ``-m
 memory_profiler`` in the command line, the ``precision`` parameter is ignored.
 
-Time-based memory usage 
+Time-based memory usage
 ==========================
 Sometimes it is useful to have full memory usage reports as a function of
 time (not line-by-line) of external processes (be it Python scripts or not).
@@ -131,14 +131,14 @@ e.g. `mprof run -h`.
 In the case of a Python script, using the previous command does not
 give you any information on which function is executed at a given
 time. Depending on the case, it can be difficult to identify the part
-of the code that is causing the highest memory usage. 
+of the code that is causing the highest memory usage.
 
 Adding the `profile` decorator to a function and running the Python
-script with 
+script with
 
     mprof run <script>
 
-will record timestamps when entering/leaving the profiled function. Runnning
+will record timestamps when entering/leaving the profiled function. Running
 
     mprof plot
 
@@ -152,9 +152,9 @@ A discussion of these capabilities can be found `here <http://fa.bianp.net/blog/
 
 .. warning:: If your Python file imports the memory profiler `from memory_profiler import profile` these timestamps will not be recorded. Comment out the import, leave your functions decorated, and re-run.
 
-The available commands for `mprof` are: 
+The available commands for `mprof` are:
 
-  - ``mprof run``: running an executable, recording memory usage  
+  - ``mprof run``: running an executable, recording memory usage
   - ``mprof plot``: plotting one the recorded memory usage (by default,
     the last one)
   - ``mprof list``: listing all recorded memory usage files in a
@@ -162,6 +162,39 @@ The available commands for `mprof` are:
   - ``mprof clean``: removing all recorded memory usage files.
   - ``mprof rm``: removing specific recorded memory usage files
 
+Tracking forked child processes
+===============================
+In a multiprocessing context the main process will spawn child processes whose
+system resources are allocated separately from the parent process. This can
+lead to an inaccurate report of memory usage since by default only the parent
+process is being tracked. The ``mprof`` utility provides two mechanisms to
+track the usage of child processes: sum the memory of all children to the
+parent's usage and track each child individual.
+
+To create a report that combines memory usage of all the children and the
+parent, use the ``include_children`` flag in either the ``profile`` decorator or
+ass a command line argument to ``mprof``::
+
+    mprof run --include-children <script>
+
+The second method tracks each child independently of the main process,
+serializing child rows by index to the output stream. Use the ``multiprocess``
+flag and plot as follows::
+
+    mprof run --multiprocess <script>
+    mprof plot
+
+This will create a plot using matplotlib similar to this:
+
+.. image:: https://cloud.githubusercontent.com/assets/745966/24075879/2e85b43a-0bfa-11e7-8dfe-654320dbd2ce.png
+    :target: https://github.com/fabianp/memory_profiler/pull/134
+    :height: 350px
+
+You can combine both the ``include_children`` and ``multiprocess`` flags to show
+the total memory of the program as well as each child individually. If using
+the API directly, note that the return from ``memory_usage`` will include the
+child memory in a nested list along with the main process memory.
+
 Setting debugger breakpoints
 =============================
 It is possible to set breakpoints depending on the amount of memory used.
@@ -260,7 +293,7 @@ LogFile of memory profiler module.
     >>> import sys
     >>> sys.stdout = LogFile('memory_profile_log')
 
-``Customised reporting:``
+``Customized reporting:``
 
 Sending everything to the log file while running the memory_profiler
 could be cumbersome and one can choose only entries with increments
@@ -412,6 +445,8 @@ cleanup.
 
 `Dmitriy Novozhilov <https://github.com/demiurg906>`_ and `Sergei Lebedev <https://github.com/superbobry>`_ added support for `tracemalloc <https://docs.python.org/3/library/tracemalloc.html>`_.
 
+`Benjamin Bengfort <https://github.com/bbengfort>`_ added support for tracking the usage of individual child processes and plotting them.
+
 =========
  License
 =========
diff --git a/examples/multiprocessing_example.py b/examples/multiprocessing_example.py
new file mode 100644
index 0000000..9818ede
--- /dev/null
+++ b/examples/multiprocessing_example.py
@@ -0,0 +1,55 @@
+"""
+An undecorated example of a script that allocates memory in multiprocessing
+workers to demonstrate the use of memory_profiler with multiple processes.
+
+Run this script with mprof run -M python multiprocessing_example.py
+You can then visualize the usage with mprof plot.
+"""
+
+import time
+import multiprocessing as mp
+
+# Big numbers
+X6 = 10 ** 6
+X7 = 10 ** 7
+
+
+def worker(num, wait, amt=X6):
+    """
+    A function that allocates memory over time.
+    """
+    frame = []
+
+    for idx in range(num):
+        frame.extend([1] * amt)
+        time.sleep(wait)
+
+    del frame
+
+
+def main_sequential():
+    """
+    A sequential version of the work, where one worker is called at a time.
+    """
+    worker(5, 5, X6)
+    worker(5, 2, X7)
+    worker(5, 5, X6)
+    worker(5, 2, X7)
+
+
+def main_multiproc():
+    """
+    A multiprocessing version of the work, where workers work in their own
+    child processes and are collected by the master process.
+    """
+    pool    = mp.Pool(processes=4)
+    tasks   = [
+        pool.apply_async(worker, args) for args in
+        [(5, 5, X6), (5, 2, X7), (5, 5, X6), (5, 2, X7)]
+    ]
+
+    results = [p.get() for p in tasks]
+
+
+if __name__ == '__main__':
+    main_multiproc()
diff --git a/memory_profiler.py b/memory_profiler.py
index afe8709..3c1e18b 100644
--- a/memory_profiler.py
+++ b/memory_profiler.py
@@ -86,6 +86,38 @@ def _repr_pretty_(self, p, cycle):
         p.text(u'<MemitResult : ' + msg + u'>')
 
 
+def _get_child_memory(process, meminfo_attr=None):
+    """
+    Returns a generator that yields memory for all child processes.
+    """
+    if not has_psutil:
+        raise NotImplementedError((
+            "The psutil module is required to monitor the "
+            "memory usage of child processes."
+        ))
+
+    # Convert a pid to a process
+    if isinstance(process, int):
+        if process == -1: process = os.getpid()
+        process = psutil.Process(process)
+
+    if not meminfo_attr:
+        # Use the psutil 2.0 attr if the older version isn't passed in.
+        meminfo_attr = 'memory_info' if hasattr(process, 'memory_info') else 'get_memory_info'
+
+    # Select the psutil function get the children similar to how we selected
+    # the memory_info attr (a change from excepting the AttributeError).
+    children_attr = 'children' if hasattr(process, 'children') else 'get_children'
+
+    # Loop over the child processes and yield their memory
+    try:
+        for child in getattr(process, children_attr)(recursive=True):
+            yield getattr(child, meminfo_attr)()[0] / _TWO_20
+    except psutil.NoSuchProcess:
+        # https://github.com/fabianp/memory_profiler/issues/71
+        yield 0.0
+
+
 def _get_memory(pid, backend, timestamps=False, include_children=False, filename=None):
     # .. low function to get memory consumption ..
     if pid == -1:
@@ -111,12 +143,7 @@ def ps_util_tool():
                 else 'get_memory_info'
             mem = getattr(process, meminfo_attr)()[0] / _TWO_20
             if include_children:
-                try:
-                    for p in process.children(recursive=True):
-                        mem += getattr(p, meminfo_attr)()[0] / _TWO_20
-                except psutil.NoSuchProcess:
-                    # https://github.com/fabianp/memory_profiler/issues/71
-                    pass
+                mem +=  sum(_get_child_memory(process, meminfo_attr))
             if timestamps:
                 return mem, time.time()
             else:
@@ -128,9 +155,11 @@ def ps_util_tool():
     def posix_tool():
         # .. scary stuff ..
         if include_children:
-            raise NotImplementedError('The psutil module is required when to'
-                                      ' monitor memory usage of children'
-                                      ' processes')
+            raise NotImplementedError((
+                "The psutil module is required to monitor the "
+                "memory usage of child processes."
+            ))
+
         warnings.warn("psutil module not found. memory_profiler will be slow")
         # ..
         # .. memory usage in MiB ..
@@ -211,8 +240,8 @@ def run(self):
 
 
 def memory_usage(proc=-1, interval=.1, timeout=None, timestamps=False,
-                 include_children=False, max_usage=False, retval=False,
-                 stream=None, backend=None):
+                 include_children=False, multiprocess=False, max_usage=False,
+                 retval=False, stream=None, backend=None):
     """
     Return the memory usage of a process or piece of code
 
@@ -243,6 +272,12 @@ def memory_usage(proc=-1, interval=.1, timeout=None, timestamps=False,
     timestamps : bool, optional
         if True, timestamps of memory usage measurement are collected as well.
 
+    include_children : bool, optional
+        if True, sum the memory of all forked processes as well
+
+    multiprocess : bool, optional
+        if True, track the memory usage of all forked processes.
+
     stream : File
         if stream is a File opened with write access, then results are written
         to this file instead of stored in memory and returned at the end of
@@ -314,9 +349,22 @@ def memory_usage(proc=-1, interval=.1, timeout=None, timestamps=False,
                 mem_usage = _get_memory(
                     proc.pid, backend, timestamps=timestamps,
                     include_children=include_children)
+
                 if stream is not None:
                     stream.write("MEM {0:.6f} {1:.4f}\n".format(*mem_usage))
+
+                    # Write children to the stream file
+                    if multiprocess:
+                        for idx, chldmem in enumerate(_get_child_memory(proc.pid)):
+                            stream.write("CHLD {0} {1:.6f} {2:.4f}\n".format(idx, chldmem, time.time()))
                 else:
+                    # Create a nested list with the child memory
+                    if multiprocess:
+                        mem_usage = [mem_usage]
+                        for chldmem in _get_child_memory(proc.pid):
+                            mem_usage.append(chldmem)
+
+                    # Append the memory usage to the return value
                     ret.append(mem_usage)
             else:
                 ret = max(ret,
@@ -348,7 +396,19 @@ def memory_usage(proc=-1, interval=.1, timeout=None, timestamps=False,
                     include_children=include_children)
                 if stream is not None:
                     stream.write("MEM {0:.6f} {1:.4f}\n".format(*mem_usage))
+
+                    # Write children to the stream file
+                    if multiprocess:
+                        for idx, chldmem in enumerate(_get_child_memory(proc.pid)):
+                            stream.write("CHLD {0} {1:.6f} {2:.4f}\n".format(idx, chldmem, time.time()))
                 else:
+                    # Create a nested list with the child memory
+                    if multiprocess:
+                        mem_usage = [mem_usage]
+                        for chldmem in _get_child_memory(proc.pid):
+                            mem_usage.append(chldmem)
+
+                    # Append the memory usage to the return value
                     ret.append(mem_usage)
             else:
                 ret = max([ret,
diff --git a/mprof b/mprof
index 192ba10..c13da08 100755
--- a/mprof
+++ b/mprof
@@ -9,6 +9,7 @@ import copy
 import time
 import math
 
+from collections import defaultdict
 from optparse import OptionParser, OptionValueError
 
 import memory_profiler as mp
@@ -195,6 +196,9 @@ def run_action():
     parser.add_option("--include-children", "-C", dest="include_children",
                       default=False, action="store_true",
                       help="""Monitors forked processes as well (sum up all process memory)""")
+    parser.add_option("--multiprocess", "-M", dest="multiprocess",
+                      default=False, action="store_true",
+                      help="""Monitors forked processes creating individual plots for each child""")
 
     (options, args) = parser.parse_args()
 
@@ -231,7 +235,8 @@ def run_action():
     with open(mprofile_output, "a") as f:
         f.write("CMDLINE {0}\n".format(cmd_line))
         mp.memory_usage(proc=p, interval=options.interval, timestamps=True,
-                        include_children=options.include_children, stream=f)
+                        include_children=options.include_children,
+                        multiprocess=options.multiprocess, stream=f)
 
 
 def add_brackets(xloc, yloc, xshift=0, color="r", label=None, options=None):
@@ -299,6 +304,7 @@ def read_mprofile_file(filename):
     func_ts = {}
     mem_usage = []
     timestamp = []
+    children  = defaultdict(list)
     cmd_line = None
     f = open(filename, "r")
     for l in f:
@@ -319,6 +325,13 @@ def read_mprofile_file(filename):
                        float(mem_start), float(mem_end)])
             func_ts[f_name] = ts
 
+        elif field == "CHLD":
+            values = value.split(' ')
+            chldnum = values[0]
+            children[chldnum].append(
+                (float(values[1]), float(values[2]))
+            )
+
         elif field == "CMDLINE":
             cmd_line = value
         else:
@@ -327,10 +340,10 @@ def read_mprofile_file(filename):
 
     return {"mem_usage": mem_usage, "timestamp": timestamp,
             "func_timestamp": func_ts, 'filename': filename,
-            'cmd_line': cmd_line}
+            'cmd_line': cmd_line, 'children': children}
 
 
-def plot_file(filename, index=0, timestamps=True, options=None):
+def plot_file(filename, index=0, timestamps=True, children=True, options=None):
     try:
         import pylab as pl
     except ImportError:
@@ -351,6 +364,7 @@ def plot_file(filename, index=0, timestamps=True, options=None):
     ts = mprofile['func_timestamp']
     t = mprofile['timestamp']
     mem = mprofile['mem_usage']
+    chld = mprofile['children']
 
     if len(ts) > 0:
         for values in ts.values():
@@ -384,6 +398,28 @@ def plot_file(filename, index=0, timestamps=True, options=None):
     bottom += 0.001
     top -= 0.001
 
+    # plot children, if any
+    if len(chld) > 0 and children:
+        cmpoint = (0,0) # maximal child memory
+
+        for idx, (proc, data) in enumerate(chld.items()):
+            # Create the numpy arrays from the series data
+            cts  = np.asarray([item[1] for item in data]) - global_start
+            cmem = np.asarray([item[0] for item in data])
+
+            # Plot the line to the figure
+            pl.plot(cts, cmem, "+-"  + mem_line_colors[idx+1 % len(mem_line_colors)],
+                     label="child {}".format(proc))
+
+            # Detect the maximal child memory point
+            cmax_mem = cmem.max()
+            if cmax_mem > cmpoint[1]:
+                cmpoint = (cts[cmem.argmax()], cmax_mem)
+
+        # Add the marker lines for the maximal child memory usage
+        pl.vlines(cmpoint[0], pl.ylim()[0]+0.001, pl.ylim()[1] - 0.001, 'r', '--')
+        pl.hlines(cmpoint[1], pl.xlim()[0]+0.001, pl.xlim()[1] - 0.001, 'r', '--')
+
     # plot timestamps, if any
     if len(ts) > 0 and timestamps:
         func_num = 0