diff --git a/doc/source/whatsnew/v1.2.0.rst b/doc/source/whatsnew/v1.2.0.rst
index 3992e697db7e4..0aa82fb936a95 100644
--- a/doc/source/whatsnew/v1.2.0.rst
+++ b/doc/source/whatsnew/v1.2.0.rst
@@ -336,7 +336,7 @@ Groupby/resample/rolling
 - Bug in :meth:`DataFrameGroupby.tshift` failing to raise ``ValueError`` when a frequency cannot be inferred for the index of a group (:issue:`35937`)
 - Bug in :meth:`DataFrame.groupby` does not always maintain column index name for ``any``, ``all``, ``bfill``, ``ffill``, ``shift`` (:issue:`29764`)
 - Bug in :meth:`DataFrameGroupBy.apply` raising error with ``np.nan`` group(s) when ``dropna=False`` (:issue:`35889`)
--
+- Bug in :meth:`Rolling.sum()` returned wrong values when dtypes where mixed between float and integer and axis was equal to one (:issue:`20649`, :issue:`35596`)
 
 Reshaping
 ^^^^^^^^^
diff --git a/pandas/core/window/rolling.py b/pandas/core/window/rolling.py
index 21a7164411fb7..06c3ad23f904f 100644
--- a/pandas/core/window/rolling.py
+++ b/pandas/core/window/rolling.py
@@ -243,7 +243,13 @@ def _create_data(self, obj: FrameOrSeries) -> FrameOrSeries:
         if self.on is not None and not isinstance(self.on, Index):
             if obj.ndim == 2:
                 obj = obj.reindex(columns=obj.columns.difference([self.on]), copy=False)
-
+        if self.axis == 1:
+            # GH: 20649 in case of mixed dtype and axis=1 we have to convert everything
+            # to float to calculate the complete row at once. We exclude all non-numeric
+            # dtypes.
+            obj = obj.select_dtypes(include=["integer", "float"], exclude=["timedelta"])
+            obj = obj.astype("float64", copy=False)
+            obj._mgr = obj._mgr.consolidate()
         return obj
 
     def _gotitem(self, key, ndim, subset=None):
diff --git a/pandas/tests/window/test_rolling.py b/pandas/tests/window/test_rolling.py
index 88afcec0f7bf4..4dfa0287bbb03 100644
--- a/pandas/tests/window/test_rolling.py
+++ b/pandas/tests/window/test_rolling.py
@@ -771,3 +771,51 @@ def test_rolling_numerical_too_large_numbers():
         index=dates,
     )
     tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    ("func", "value"),
+    [("sum", 2.0), ("max", 1.0), ("min", 1.0), ("mean", 1.0), ("median", 1.0)],
+)
+def test_rolling_mixed_dtypes_axis_1(func, value):
+    # GH: 20649
+    df = pd.DataFrame(1, index=[1, 2], columns=["a", "b", "c"])
+    df["c"] = 1.0
+    result = getattr(df.rolling(window=2, min_periods=1, axis=1), func)()
+    expected = pd.DataFrame(
+        {"a": [1.0, 1.0], "b": [value, value], "c": [value, value]}, index=[1, 2]
+    )
+    tm.assert_frame_equal(result, expected)
+
+
+def test_rolling_axis_one_with_nan():
+    # GH: 35596
+    df = pd.DataFrame(
+        [
+            [0, 1, 2, 4, np.nan, np.nan, np.nan],
+            [0, 1, 2, np.nan, np.nan, np.nan, np.nan],
+            [0, 2, 2, np.nan, 2, np.nan, 1],
+        ]
+    )
+    result = df.rolling(window=7, min_periods=1, axis="columns").sum()
+    expected = pd.DataFrame(
+        [
+            [0.0, 1.0, 3.0, 7.0, 7.0, 7.0, 7.0],
+            [0.0, 1.0, 3.0, 3.0, 3.0, 3.0, 3.0],
+            [0.0, 2.0, 4.0, 4.0, 6.0, 6.0, 7.0],
+        ]
+    )
+    tm.assert_frame_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "value",
+    ["test", pd.to_datetime("2019-12-31"), pd.to_timedelta("1 days 06:05:01.00003")],
+)
+def test_rolling_axis_1_non_numeric_dtypes(value):
+    # GH: 20649
+    df = pd.DataFrame({"a": [1, 2]})
+    df["b"] = value
+    result = df.rolling(window=2, min_periods=1, axis=1).sum()
+    expected = pd.DataFrame({"a": [1.0, 2.0]})
+    tm.assert_frame_equal(result, expected)