diff --git a/python/pyspark/serializers.py b/python/pyspark/serializers.py
index 4c16b5fc26f3..82abf1947c81 100644
--- a/python/pyspark/serializers.py
+++ b/python/pyspark/serializers.py
@@ -216,9 +216,10 @@ def _create_batch(series, timezone):
     :param timezone: A timezone to respect when handling timestamp values
     :return: Arrow RecordBatch
     """
-
-    from pyspark.sql.types import _check_series_convert_timestamps_internal
+    import decimal
+    from distutils.version import LooseVersion
     import pyarrow as pa
+    from pyspark.sql.types import _check_series_convert_timestamps_internal
     # Make input conform to [(series1, type1), (series2, type2), ...]
     if not isinstance(series, (list, tuple)) or \
             (len(series) == 2 and isinstance(series[1], pa.DataType)):
@@ -236,6 +237,11 @@ def create_array(s, t):
             # TODO: need decode before converting to Arrow in Python 2
             return pa.Array.from_pandas(s.apply(
                 lambda v: v.decode("utf-8") if isinstance(v, str) else v), mask=mask, type=t)
+        elif t is not None and pa.types.is_decimal(t) and \
+                LooseVersion("0.9.0") <= LooseVersion(pa.__version__) < LooseVersion("0.10.0"):
+            # TODO: see ARROW-2432. Remove when the minimum PyArrow version becomes 0.10.0.
+            return pa.Array.from_pandas(s.apply(
+                lambda v: decimal.Decimal('NaN') if v is None else v), mask=mask, type=t)
         return pa.Array.from_pandas(s, mask=mask, type=t)
 
     arrs = [create_array(s, t) for s, t in series]