In view of the traditional A* algorithm has the disadvantages of large search range,many turning points and poor real-time performance in AGV path planning,based on the A* algorithm,the heuristic function was improved by establishing grid map to remove redundant nodes and improve obstacle avoidance security.Aiming at the dynamic path planning problem of AGV in complex environment,the improved A* algorithm and the dynamic window algorithm were fused to plan an optimal path with real-time performance.Through the simulation experiments,the effectiveness and feasibility of the improved algorithm were verified,and the path optimization was realized.The experiment was carried out through the robot operating system.The results show that the AGV path planning is reasonable and meets the practical application requirements.