diff --git a/docs/releases/0.3-NOTES.md b/docs/releases/0.3-NOTES.md
new file mode 100644
index 000000000..1dece9b47
--- /dev/null
+++ b/docs/releases/0.3-NOTES.md
@@ -0,0 +1,31 @@
+## Release notes for firmament v0.3
+
+# Significant changes
+
+* Support for Pod Affinity and Pod Anti-Affinity policy.
+* Please refer the k8s documentation [here](https://kubernetes.io/docs/concepts/configuration/assign-pod-node/) on Pod Affinity and Pod Anti-Affinity support.
+* Also the firmament/poseidon design documents for pod affinity and anti-affinity can be found [here](https://docs.google.com/document/d/1C68E1Thw7i9NGlHFwmnkgWmNSUK3z0oHdH2mLMK9Eyw/edit)
+
+* Bug fixes.
+
+# Required Actions
+
+To enable the new CPU-MEM cost model ```--flow_scheduling_cost_model```
+flag should be set to '10'
+like shown below
+
+```--flow_scheduling_cost_model=10```
+
+This fix is based above the CPU-MEM cost model.
+This feature currently are not supported in other cost model.
+
+# Highlighted changes
+
+Pod Affinity and Pod Anti-Affinity support.
+
+# Full change list
+
+* [Pod Affinity and Anti-affinity feature support](https://github.com/Huawei-PaaS/firmament/pull/4)
+* [Support for GT an LT operators for node affinity](https://github.com/Huawei-PaaS/firmament/pull/5)
+* [Improved the static resource calculation](https://github.com/Huawei-PaaS/firmament/pull/6)
+* [Update README.md](https://github.com/Huawei-PaaS/firmament/pull/2)
\ No newline at end of file