diff --git a/doc/source/whatsnew/v0.25.1.rst b/doc/source/whatsnew/v0.25.1.rst
index eb60272246ebb..c0484f9217396 100644
--- a/doc/source/whatsnew/v0.25.1.rst
+++ b/doc/source/whatsnew/v0.25.1.rst
@@ -128,7 +128,7 @@ Groupby/resample/rolling
 Reshaping
 ^^^^^^^^^
 
--
+- Bug in :meth:`DataFrame.crosstab` when margins set to True and normalize is not False, an error is raised. (:issue:`27500`)
 -
 -
 
diff --git a/pandas/core/reshape/pivot.py b/pandas/core/reshape/pivot.py
index 79716520f6654..1112f0393908a 100644
--- a/pandas/core/reshape/pivot.py
+++ b/pandas/core/reshape/pivot.py
@@ -611,13 +611,17 @@ def _normalize(table, normalize, margins, margins_name="All"):
         table = table.fillna(0)
 
     elif margins is True:
-
+        # keep index and column of pivoted table
+        table_index = table.index
+        table_columns = table.columns
+        # drop margins created in pivot_table and only keep the core
         column_margin = table.loc[:, margins_name].drop(margins_name)
-        index_margin = table.loc[margins_name, :].drop(margins_name)
+        # separate cases between multiindex and index
+        if isinstance(table_index, MultiIndex):
+            index_margin = table.loc[margins_name, :].drop(margins_name, axis=1)
+        else:
+            index_margin = table.loc[margins_name, :].drop(margins_name)
         table = table.drop(margins_name, axis=1).drop(margins_name)
-        # to keep index and columns names
-        table_index_names = table.index.names
-        table_columns_names = table.columns.names
 
         # Normalize core
         table = _normalize(table, normalize=normalize, margins=False)
@@ -627,11 +631,19 @@ def _normalize(table, normalize, margins, margins_name="All"):
             column_margin = column_margin / column_margin.sum()
             table = concat([table, column_margin], axis=1)
             table = table.fillna(0)
+            table.columns = table_columns
 
         elif normalize == "index":
-            index_margin = index_margin / index_margin.sum()
+            # index_margin is a dataframe, and use a hacky way: sum(axis=1)[0]
+            # to get the normalized result, and use sum() instead for series
+            if isinstance(index_margin, ABCDataFrame):
+                sum_index_margin = index_margin.sum(axis=1)[0]
+            else:
+                sum_index_margin = index_margin.sum()
+            index_margin = index_margin / sum_index_margin
             table = table.append(index_margin)
             table = table.fillna(0)
+            table.index = table_index
 
         elif normalize == "all" or normalize is True:
             column_margin = column_margin / column_margin.sum()
@@ -641,13 +653,12 @@ def _normalize(table, normalize, margins, margins_name="All"):
             table = table.append(index_margin)
 
             table = table.fillna(0)
+            table.index = table_index
+            table.columns = table_columns
 
         else:
             raise ValueError("Not a valid normalize argument")
 
-        table.index.names = table_index_names
-        table.columns.names = table_columns_names
-
     else:
         raise ValueError("Not a valid margins argument")
 
diff --git a/pandas/tests/reshape/test_pivot.py b/pandas/tests/reshape/test_pivot.py
index be82e7f595f8c..ac12165aa0fd1 100644
--- a/pandas/tests/reshape/test_pivot.py
+++ b/pandas/tests/reshape/test_pivot.py
@@ -2447,3 +2447,38 @@ def test_crosstab_unsorted_order(self):
             [[1, 0, 0], [0, 1, 0], [0, 0, 1]], index=e_idx, columns=e_columns
         )
         tm.assert_frame_equal(result, expected)
+
+    def test_margin_normalize(self):
+        df = pd.DataFrame({"A": ["foo", "foo", "foo", "foo", "foo",
+                                 "bar", "bar", "bar", "bar"],
+                           "B": ["one", "one", "one", "two", "two",
+                                 "one", "one", "two", "two"],
+                           "C": ["small", "large", "large", "small",
+                                 "small", "large", "small", "small",
+                                 "large"],
+                           "D": [1, 2, 2, 3, 3, 4, 5, 6, 7],
+                           "E": [2, 4, 5, 5, 6, 6, 8, 9, 9]})
+        # normalize on index
+        result = pd.crosstab([df.A, df.B], df.C, margins=True, margins_name='Sub-Total',
+                             normalize=0)
+        expected = pd.DataFrame([[0.5, 0.5], [0.5, 0.5],
+                                 [0.666667, 0.333333], [0, 1],
+                                 [0.444444, 0.555556]])
+        expected.index = MultiIndex(levels=[['Sub-Total', 'bar', 'foo'],
+                                            ['', 'one', 'two']],
+                                    codes=[[1, 1, 2, 2, 0], [1, 2, 1, 2, 0]],
+                                    names=['A', 'B'])
+        expected.columns = Index(['large', 'small'], dtype='object', name='C')
+        tm.assert_frame_equal(result, expected)
+
+        # normalize on columns
+        result = pd.crosstab([df.A, df.B], df.C, margins=True, margins_name='Sub-Total',
+                             normalize=1)
+        expected = pd.DataFrame([[0.25, 0.2, 0.222222], [0.25, 0.2, 0.222222],
+                                 [0.5, 0.2, 0.333333], [0, 0.4, 0.222222]])
+        expected.columns = Index(['large', 'small', 'Sub-Total'], dtype='object',
+                                 name='C')
+        expected.index = MultiIndex(levels=[['bar', 'foo'], ['one', 'two']],
+                                    codes=[[1, 1, 2, 2], [1, 2, 1, 2]],
+                                    names=['A, B'])
+        tm.assert_frame_equal(result, expected)