diff --git a/opentelemetry-sdk/tests/trace/test_span_processor.py b/opentelemetry-sdk/tests/trace/test_span_processor.py
index b8568fc7a1b..c672d4ce102 100644
--- a/opentelemetry-sdk/tests/trace/test_span_processor.py
+++ b/opentelemetry-sdk/tests/trace/test_span_processor.py
@@ -16,10 +16,13 @@
 import time
 import typing
 import unittest
+from platform import python_implementation, system
 from threading import Event
 from typing import Optional
 from unittest import mock
 
+from pytest import mark
+
 from opentelemetry import trace as trace_api
 from opentelemetry.context import Context
 from opentelemetry.sdk import trace
@@ -266,6 +269,10 @@ def create_multi_span_processor(
     ) -> trace.ConcurrentMultiSpanProcessor:
         return trace.ConcurrentMultiSpanProcessor(3)
 
+    @mark.skipif(
+        python_implementation() == "PyPy" and system() == "Windows",
+        reason="This test randomly fails in Windows with PyPy",
+    )
     def test_force_flush_late_by_timeout(self):
         multi_processor = trace.ConcurrentMultiSpanProcessor(5)
         wait_event = Event()