diff --git a/pandas/io/parquet.py b/pandas/io/parquet.py
index aec31f40f8570..3d8346aeb97a2 100644
--- a/pandas/io/parquet.py
+++ b/pandas/io/parquet.py
@@ -24,7 +24,9 @@
 
 from pandas import (
     DataFrame,
+    Index,
     MultiIndex,
+    RangeIndex,
     arrays,
     get_option,
 )
@@ -250,14 +252,37 @@ def read(
             if dtype_backend == "pandas":
                 result = pa_table.to_pandas(**to_pandas_kwargs)
             elif dtype_backend == "pyarrow":
-                result = DataFrame(
-                    {
-                        col_name: arrays.ArrowExtensionArray(pa_col)
-                        for col_name, pa_col in zip(
-                            pa_table.column_names, pa_table.itercolumns()
-                        )
-                    }
-                )
+                index_columns = pa_table.schema.pandas_metadata.get("index_columns", [])
+                result_dc = {
+                    col_name: arrays.ArrowExtensionArray(pa_col)
+                    for col_name, pa_col in zip(
+                        pa_table.column_names, pa_table.itercolumns()
+                    )
+                }
+                idx: Index | None
+                if len(index_columns) == 0:
+                    idx = None
+                elif len(index_columns) == 1 and isinstance(index_columns[0], dict):
+                    params = index_columns[0]
+                    idx = RangeIndex(
+                        params.get("start"),
+                        params.get("stop"),
+                        params.get("step"),
+                        name=params.get("name"),
+                    )
+
+                else:
+                    index_data = [
+                        result_dc.pop(index_col) for index_col in index_columns
+                    ]
+                    if len(index_data) == 1:
+                        name = index_columns[0]
+                        if isinstance(name, str) and name.startswith("__index_level_"):
+                            name = None
+                        idx = Index(index_data[0], name=name)
+                    else:
+                        idx = MultiIndex.from_arrays(index_data, names=index_columns)
+                result = DataFrame(result_dc, index=idx)
             if manager == "array":
                 result = result._as_manager("array", copy=False)
             return result
diff --git a/pandas/tests/io/test_parquet.py b/pandas/tests/io/test_parquet.py
index 353dc4f1cbd8a..b8d02fc04f90e 100644
--- a/pandas/tests/io/test_parquet.py
+++ b/pandas/tests/io/test_parquet.py
@@ -18,6 +18,7 @@
 import pandas.util._test_decorators as td
 
 import pandas as pd
+from pandas import RangeIndex
 import pandas._testing as tm
 from pandas.util.version import Version
 
@@ -1225,3 +1226,38 @@ def test_bytes_file_name(self, engine):
 
             result = read_parquet(path, engine=engine)
         tm.assert_frame_equal(result, df)
+
+    @pytest.mark.parametrize("index", ["A", ["A", "B"]])
+    def test_pyarrow_backed_df_index(self, index, pa):
+        # GH#48944
+        obj = pd.DataFrame(data={"A": [0, 1], "B": [1, 0], "C": 1})
+        df = obj.set_index(index)
+        with tm.ensure_clean("test.parquet") as path:
+            with open(path.encode(), "wb") as f:
+                df.to_parquet(f)
+
+            with pd.option_context("mode.dtype_backend", "pyarrow"):
+                result = read_parquet(path, engine="pyarrow")
+        expected = obj.astype("int64[pyarrow]").set_index(index)
+        tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("name", [None, "test"])
+    @pytest.mark.parametrize("index", [True, False, None])
+    def test_pyarrow_backed_df_range_index(self, pa, index, name):
+        # GH#48944
+        df = pd.DataFrame(
+            data={"A": [0, 1], "B": [1, 0]},
+            index=RangeIndex(start=100, stop=102, name=name),
+        )
+        with tm.ensure_clean("test.parquet") as path:
+            with open(path.encode(), "wb") as f:
+                df.to_parquet(f, index=index)
+
+            with pd.option_context("mode.dtype_backend", "pyarrow"):
+                result = read_parquet(path, engine="pyarrow")
+        expected = df.astype("int64[pyarrow]")
+        if index is False:
+            expected = expected.reset_index(drop=True)
+        elif index:
+            expected.index = pd.Index([100, 101], dtype="int64[pyarrow]", name=name)
+        tm.assert_frame_equal(result, expected)