diff --git a/README.md b/README.md
index ac7e719..6442018 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,23 @@
 # TetrisRL
 - Implement a reinforcement learning model which can play Tetris
 
+# Tetris Enviroment
+- 10 actions
+    - 0: don't move
+    - 1: shift left 1 block
+    - 2: shift left 2 block
+    - 3: shift left 3 block
+    - 4: shift right 1 block
+    - 5: shift right 2 block
+    - 6: shift right 3 block
+    - 7: rotate once
+    - 8: rotate twice
+    - 9: rotate three times
+- return
+    - pixel(10*20)
+    - block_id
+    - block_location(x, y)
+
 # Version 0
 - Policy Gradient Algorithm
 - Reward Delay