diff --git a/src/awkward/operations/ak_to_dataframe.py b/src/awkward/operations/ak_to_dataframe.py
index 46e4206ad5..023d23f8de 100644
--- a/src/awkward/operations/ak_to_dataframe.py
+++ b/src/awkward/operations/ak_to_dataframe.py
@@ -155,6 +155,35 @@ def _impl(array, how, levelname, anonymous):
     conda install pandas"""
         ) from err
 
+    def _merge_multi_indexed_dfs(
+        df1: pandas.DataFrame, df2: pandas.DataFrame, how: str
+    ) -> pandas.DataFrame:
+        """Merge two DataFrames after aligning their MultiIndex levels."""
+        # Extra levels. At most one of these will be non-empty.
+        df1_extra = tuple(n for n in df1.index.names if n not in df2.index.names)
+        df2_extra = tuple(n for n in df2.index.names if n not in df1.index.names)
+
+        # Reset extra levels to columns
+        if df1_extra:
+            df1 = df1.reset_index(level=df1_extra)
+        if df2_extra:
+            df2 = df2.reset_index(level=df2_extra)
+
+        # Merge on aligned indices
+        merged = pandas.merge(df1, df2, left_index=True, right_index=True, how=how)
+
+        # Fill missing values in extra levels and convert to int
+        for n in df1_extra:
+            merged[n] = merged[n].fillna(0).astype(int)
+        for n in df2_extra:
+            merged[n] = merged[n].fillna(0).astype(int)
+
+        # Set index back with extra levels
+        if df1_extra or df2_extra:
+            merged = merged.set_index(list(df1_extra) + list(df2_extra), append=True)
+
+        return merged
+
     if how is not None:
         out = None
         for df in to_dataframe(
@@ -163,7 +192,7 @@ def _impl(array, how, levelname, anonymous):
             if out is None:
                 out = df
             else:
-                out = pandas.merge(out, df, how=how, left_index=True, right_index=True)
+                out = _merge_multi_indexed_dfs(out, df, how)
         return out
 
     def recurse(layout, row_arrays, col_names):
diff --git a/tests/test_0080_flatpandas_multiindex_rows_and_columns.py b/tests/test_0080_flatpandas_multiindex_rows_and_columns.py
index 5ef2eea3a1..a9ed197439 100644
--- a/tests/test_0080_flatpandas_multiindex_rows_and_columns.py
+++ b/tests/test_0080_flatpandas_multiindex_rows_and_columns.py
@@ -140,40 +140,40 @@ def regularize(data):
         json.loads(ak.operations.to_dataframe(array, how="outer").to_json())
     ) == {
         "x": {
-            (0, 0, 0, None): 0.0,
-            (0, 0, 1, 0.0): 1.1,
-            (0, 0, 2, 0.0): 2.2,
-            (0, 0, 2, 1.0): 2.2,
-            (0, 2, 0, 0.0): 3.3,
-            (0, 2, 0, 1.0): 3.3,
-            (0, 2, 0, 2.0): 3.3,
-            (0, 2, 1, 0.0): 4.4,
-            (0, 2, 1, 1.0): 4.4,
-            (0, 2, 1, 2.0): 4.4,
-            (0, 2, 1, 3.0): 4.4,
-            (2, 0, 0, 0.0): 5.5,
-            (2, 0, 0, 1.0): 5.5,
-            (2, 0, 0, 2.0): 5.5,
-            (2, 0, 0, 3.0): 5.5,
-            (2, 0, 0, 4.0): 5.5,
+            (0, 0, 0, 0): 0.0,
+            (0, 0, 1, 0): 1.1,
+            (0, 0, 2, 0): 2.2,
+            (0, 0, 2, 1): 2.2,
+            (0, 2, 0, 0): 3.3,
+            (0, 2, 0, 1): 3.3,
+            (0, 2, 0, 2): 3.3,
+            (0, 2, 1, 0): 4.4,
+            (0, 2, 1, 1): 4.4,
+            (0, 2, 1, 2): 4.4,
+            (0, 2, 1, 3): 4.4,
+            (2, 0, 0, 0): 5.5,
+            (2, 0, 0, 1): 5.5,
+            (2, 0, 0, 2): 5.5,
+            (2, 0, 0, 3): 5.5,
+            (2, 0, 0, 4): 5.5,
         },
         "y": {
-            (0, 0, 0, None): None,
-            (0, 0, 1, 0.0): 1.0,
-            (0, 0, 2, 0.0): 2.0,
-            (0, 0, 2, 1.0): 2.0,
-            (0, 2, 0, 0.0): 3.0,
-            (0, 2, 0, 1.0): 3.0,
-            (0, 2, 0, 2.0): 3.0,
-            (0, 2, 1, 0.0): 4.0,
-            (0, 2, 1, 1.0): 4.0,
-            (0, 2, 1, 2.0): 4.0,
-            (0, 2, 1, 3.0): 4.0,
-            (2, 0, 0, 0.0): 5.0,
-            (2, 0, 0, 1.0): 5.0,
-            (2, 0, 0, 2.0): 5.0,
-            (2, 0, 0, 3.0): 5.0,
-            (2, 0, 0, 4.0): 5.0,
+            (0, 0, 0, 0): None,
+            (0, 0, 1, 0): 1.0,
+            (0, 0, 2, 0): 2.0,
+            (0, 0, 2, 1): 2.0,
+            (0, 2, 0, 0): 3.0,
+            (0, 2, 0, 1): 3.0,
+            (0, 2, 0, 2): 3.0,
+            (0, 2, 1, 0): 4.0,
+            (0, 2, 1, 1): 4.0,
+            (0, 2, 1, 2): 4.0,
+            (0, 2, 1, 3): 4.0,
+            (2, 0, 0, 0): 5.0,
+            (2, 0, 0, 1): 5.0,
+            (2, 0, 0, 2): 5.0,
+            (2, 0, 0, 3): 5.0,
+            (2, 0, 0, 4): 5.0,
         },
     }
 
diff --git a/tests/test_3694_to_dataframe_align_multindex.py b/tests/test_3694_to_dataframe_align_multindex.py
new file mode 100644
index 0000000000..64364bf80c
--- /dev/null
+++ b/tests/test_3694_to_dataframe_align_multindex.py
@@ -0,0 +1,72 @@
+from __future__ import annotations
+
+import numpy as np
+import pytest
+
+import awkward as ak
+
+pd = pytest.importorskip("pandas")
+
+
+_data = {
+    "x": ["abc", "FG_12345"],
+    "y": [None, ["g1", "g2"]],
+}
+
+_expected_inner = pd.DataFrame(
+    {
+        "x": ["FG_12345", "FG_12345"],
+        "y": ["g1", "g2"],
+    },
+    index=pd.MultiIndex.from_tuples(
+        [(1, 0), (1, 1)],
+        names=["entry", "subentry"],
+    ),
+)
+
+_expected_outer = pd.DataFrame(
+    {
+        "x": ["abc", "FG_12345", "FG_12345"],
+        "y": [np.nan, "g1", "g2"],
+    },
+    index=pd.MultiIndex.from_tuples(
+        [(0, 0), (1, 0), (1, 1)],
+        names=["entry", "subentry"],
+    ),
+)
+
+_expected_left = pd.DataFrame(
+    {
+        "x": ["abc", "FG_12345", "FG_12345"],
+        "y": [np.nan, "g1", "g2"],
+    },
+    index=pd.MultiIndex.from_tuples(
+        [(0, 0), (1, 0), (1, 1)],
+        names=["entry", "subentry"],
+    ),
+)
+
+_expected_right = pd.DataFrame(
+    {
+        "x": ["FG_12345", "FG_12345"],
+        "y": ["g1", "g2"],
+    },
+    index=pd.MultiIndex.from_tuples(
+        [(1, 0), (1, 1)],
+        names=["entry", "subentry"],
+    ),
+)
+
+params = [
+    ("inner", _expected_inner),
+    ("outer", _expected_outer),
+    ("left", _expected_left),
+    ("right", _expected_right),
+]
+
+
+@pytest.mark.parametrize("how, expected", params)
+def test_merge_option(how: str, expected: pd.DataFrame) -> None:
+    a = ak.Array(_data)
+    actual = ak.to_dataframe(a, how=how)
+    pd.testing.assert_frame_equal(actual, expected)