diff --git a/pandas/core/nanops.py b/pandas/core/nanops.py
index cafd3a9915fa0..86c3c380636c9 100644
--- a/pandas/core/nanops.py
+++ b/pandas/core/nanops.py
@@ -14,7 +14,8 @@
     _get_dtype, is_any_int_dtype, is_bool_dtype, is_complex, is_complex_dtype,
     is_datetime64_dtype, is_datetime64tz_dtype, is_datetime_or_timedelta_dtype,
     is_float, is_float_dtype, is_integer, is_integer_dtype, is_numeric_dtype,
-    is_object_dtype, is_scalar, is_timedelta64_dtype)
+    is_object_dtype, is_scalar, is_timedelta64_dtype, pandas_dtype)
+from pandas.core.dtypes.dtypes import DatetimeTZDtype
 from pandas.core.dtypes.missing import isna, na_value_for_dtype, notna
 
 import pandas.core.common as com
@@ -57,7 +58,7 @@ class disallow(object):
 
     def __init__(self, *dtypes):
         super(disallow, self).__init__()
-        self.dtypes = tuple(np.dtype(dtype).type for dtype in dtypes)
+        self.dtypes = tuple(pandas_dtype(dtype).type for dtype in dtypes)
 
     def check(self, obj):
         return hasattr(obj, 'dtype') and issubclass(obj.dtype.type,
@@ -437,6 +438,7 @@ def nansum(values, axis=None, skipna=True, min_count=0, mask=None):
     return _wrap_results(the_sum, dtype)
 
 
+@disallow('M8', DatetimeTZDtype)
 @bottleneck_switch()
 def nanmean(values, axis=None, skipna=True, mask=None):
     """
diff --git a/pandas/tests/frame/test_analytics.py b/pandas/tests/frame/test_analytics.py
index f2c3f50c291c3..386e5f57617cf 100644
--- a/pandas/tests/frame/test_analytics.py
+++ b/pandas/tests/frame/test_analytics.py
@@ -794,6 +794,25 @@ def test_mean(self, float_frame_with_na, float_frame, float_string_frame):
                             check_dates=True)
         assert_stat_op_api('mean', float_frame, float_string_frame)
 
+    @pytest.mark.parametrize('tz', [None, 'UTC'])
+    def test_mean_mixed_datetime_numeric(self, tz):
+        # https://github.com/pandas-dev/pandas/issues/24752
+        df = pd.DataFrame({"A": [1, 1],
+                           "B": [pd.Timestamp('2000', tz=tz)] * 2})
+        result = df.mean()
+        expected = pd.Series([1.0], index=['A'])
+        tm.assert_series_equal(result, expected)
+
+    @pytest.mark.parametrize('tz', [None, 'UTC'])
+    def test_mean_excludeds_datetimes(self, tz):
+        # https://github.com/pandas-dev/pandas/issues/24752
+        # Our long-term desired behavior is unclear, but the behavior in
+        # 0.24.0rc1 was buggy.
+        df = pd.DataFrame({"A": [pd.Timestamp('2000', tz=tz)] * 2})
+        result = df.mean()
+        expected = pd.Series()
+        tm.assert_series_equal(result, expected)
+
     def test_product(self, float_frame_with_na, float_frame,
                      float_string_frame):
         assert_stat_op_calc('product', np.prod, float_frame_with_na)
diff --git a/pandas/tests/test_nanops.py b/pandas/tests/test_nanops.py
index 4bcd16a86e865..cf5ef6cf15eca 100644
--- a/pandas/tests/test_nanops.py
+++ b/pandas/tests/test_nanops.py
@@ -971,6 +971,9 @@ def prng(self):
 
 class TestDatetime64NaNOps(object):
     @pytest.mark.parametrize('tz', [None, 'UTC'])
+    @pytest.mark.xfail(reason="disabled")
+    # Enabling mean changes the behavior of DataFrame.mean
+    # See https://github.com/pandas-dev/pandas/issues/24752
     def test_nanmean(self, tz):
         dti = pd.date_range('2016-01-01', periods=3, tz=tz)
         expected = dti[1]