In the process of unmanned helicopter autonomous landing on ship, SLAM is the key to realize the real autonomy. This paper introduces the extended kalman filter algorithm as the basis of SLAM, describes the unmanned helicopter SLAM problem, builds up the motion equation and observation equation of unmanned helicopter, verifies the decomposition principle of EKF-SLAM algorithm by equation, studies the improved EKF-SLAM algorithm, finally verifies the algorithm by simulation experiment which can realize the navigation for unmanned helicopter autonomous landing within a certain error range.