In order to improve the efficiency and safety of dynamic programming algorithm, an optimization algorithm suitable for dynamic complex environment motion planning was designed. The jump point search (JPS) is optimized in the following 3 aspects: the state search space, the path re-planning and the path smoothness. Integrated with the dynamic window approach (DWA) trajectory planning algorithm of which the evaluation function is optimized, the real-time obstacle avoidance is realized based on the dynamic window mode. The simulation test and the comparison experiment of the A*, the JPS algorithm and the improved JPS algorithm are performed. The experimental results show the effectiveness and the safety of the proposed trajectory planning optimization algorithm.