In order to make Unmanned Aerial Vehicle (UAV) avoid unexpected static obstacles effec-tively, a method for design a controller of autonomous collision avoidance based on EKF-OSA was proposed. First,the relative motion model between an UAV and the static obstacle was built up.Then,the Extended Kalman Filter (EKF) was applied to estimate the state of the obstacle,and the rule of avoiding collision with unexpected static obstacles was presented.A guidance law based on One-Step-Ahead (OSA) optimization method was designed for collision avoidance.The simulation results veri-fied the effectiveness of the proposed approach.