diff --git a/atune.spec b/atune.spec
index b028e614fd6bec632fa7fb6add41b7b60516e0a4..38ec514127d749000bb02b8a5d13354cb08cd1b7 100755
--- a/atune.spec
+++ b/atune.spec
@@ -3,7 +3,7 @@
 Summary: AI auto tuning system
 Name: atune
 Version: 0.3
-Release: 0.3
+Release: 0.4
 License: Mulan PSL v2
 URL: https://gitee.com/openeuler/A-Tune
 Source: https://gitee.com/openeuler/A-Tune/repository/archive/v%{version}.tar.gz
@@ -52,6 +52,7 @@ Requires: python3-xgboost
 Requires: python3-flask-restful
 Requires: python3-pandas
 Requires: python3-lhsmdu
+Conflicts: atune < 0.3-0.1
 
 %description engine
 atune engine tool for manage atuned AI tuning system.
@@ -139,6 +140,9 @@ make models
 %systemd_postun_with_restart atuned.service
 
 %changelog
+* Sat Nov 28 2020 hanxinke<hanxinke@huawei.com> - 0.3-0.4
+- The engine package conflicts with atune < 0.3-0.1.
+
 * Wed Sep 9 2020 Zhipeng Xie<xiezhipeng1@huawei.com> - 0.3-0.3
 - only require prefetch_tuning on aarch64