diff --git a/doc/source/whatsnew/v2.2.0.rst b/doc/source/whatsnew/v2.2.0.rst
index 649ad37a56b35..f95c9465859cb 100644
--- a/doc/source/whatsnew/v2.2.0.rst
+++ b/doc/source/whatsnew/v2.2.0.rst
@@ -761,6 +761,7 @@ Datetimelike
 - Bug in parsing datetime strings with nanosecond resolution with non-ISO8601 formats incorrectly truncating sub-microsecond components (:issue:`56051`)
 - Bug in parsing datetime strings with sub-second resolution and trailing zeros incorrectly inferring second or millisecond resolution (:issue:`55737`)
 - Bug in the results of :func:`to_datetime` with an floating-dtype argument with ``unit`` not matching the pointwise results of :class:`Timestamp` (:issue:`56037`)
+- Fixed bug in :meth:`DataFrame.merge` not being able to join on ``datetime64`` columns of differing resolutions (:issue:`55212`)
 - Fixed regression where :func:`concat` would raise an error when concatenating ``datetime64`` columns with differing resolutions (:issue:`53641`)
 
 Timedelta
diff --git a/pandas/core/reshape/merge.py b/pandas/core/reshape/merge.py
index 4aff99dc42250..6e64a60ae4502 100644
--- a/pandas/core/reshape/merge.py
+++ b/pandas/core/reshape/merge.py
@@ -1083,18 +1083,11 @@ def _maybe_add_join_keys(
                     result_dtype = lvals.dtype
                 else:
                     key_col = Index(lvals)
+                    result_dtype = find_common_type([lvals.dtype, rvals.dtype])
+                    key_col = key_col.astype(result_dtype, copy=False)
                     if left_indexer is not None:
                         mask_left = left_indexer == -1
                         key_col = key_col.where(~mask_left, rvals)
-                    result_dtype = find_common_type([lvals.dtype, rvals.dtype])
-                    if (
-                        lvals.dtype.kind == "M"
-                        and rvals.dtype.kind == "M"
-                        and result_dtype.kind == "O"
-                    ):
-                        # TODO(non-nano) Workaround for common_type not dealing
-                        # with different resolutions
-                        result_dtype = key_col.dtype
 
                 if result._is_label_reference(name):
                     result[name] = result._constructor_sliced(
diff --git a/pandas/tests/reshape/merge/test_merge.py b/pandas/tests/reshape/merge/test_merge.py
index 2505d9163a6d2..cf3cc3eb8bdc4 100644
--- a/pandas/tests/reshape/merge/test_merge.py
+++ b/pandas/tests/reshape/merge/test_merge.py
@@ -2816,7 +2816,8 @@ def test_merge_arrow_and_numpy_dtypes(dtype):
 
 @pytest.mark.parametrize("how", ["inner", "left", "outer", "right"])
 @pytest.mark.parametrize("tz", [None, "America/Chicago"])
-def test_merge_datetime_different_resolution(tz, how):
+@pytest.mark.parametrize("unit", ["us", "ms", "s"])
+def test_merge_datetime_different_resolution(tz, how, unit):
     # https://github.com/pandas-dev/pandas/issues/53200
     vals = [
         pd.Timestamp(2023, 5, 12, tz=tz),
@@ -2826,19 +2827,31 @@ def test_merge_datetime_different_resolution(tz, how):
     df1 = DataFrame({"t": vals[:2], "a": [1.0, 2.0]})
     df1["t"] = df1["t"].dt.as_unit("ns")
     df2 = DataFrame({"t": vals[1:], "b": [1.0, 2.0]})
-    df2["t"] = df2["t"].dt.as_unit("s")
+    df2["t"] = df2["t"].dt.as_unit(unit)
 
     expected = DataFrame({"t": vals, "a": [1.0, 2.0, np.nan], "b": [np.nan, 1.0, 2.0]})
     expected["t"] = expected["t"].dt.as_unit("ns")
     if how == "inner":
-        expected = expected.iloc[[1]].reset_index(drop=True)
+        expected1 = expected.iloc[[1]].reset_index(drop=True)
+        expected2 = expected1
     elif how == "left":
-        expected = expected.iloc[[0, 1]]
+        expected1 = expected.iloc[[0, 1]]
+        expected2 = expected.iloc[[1, 2]].reset_index(drop=True)
     elif how == "right":
-        expected = expected.iloc[[1, 2]].reset_index(drop=True)
+        expected1 = expected.iloc[[1, 2]].reset_index(drop=True)
+        expected2 = expected.iloc[[0, 1]]
+    else:
+        expected1 = expected
+        expected2 = expected
 
     result = df1.merge(df2, on="t", how=how)
-    tm.assert_frame_equal(result, expected)
+    tm.assert_frame_equal(result, expected1)
+
+    # Check lower resolution to higher resolution also works
+    # GH55212
+    expected2 = expected2[["t", "b", "a"]]
+    result1 = df2.merge(df1, on="t", how=how)
+    tm.assert_frame_equal(result1, expected2)
 
 
 def test_merge_multiindex_single_level():