Abstract: In order to shorten the R&D period of AGV(automated guided vehicle) systems, reduce the cost of R&D, find out universal scheduling strategy for different maps, and improve robustness to actual scheduling tasks and malfuct, a two-stage dynamic path planning strategy is developed. Multiple AGV scheduling systems are controlled with the two-stage control strategy, and paths are acquired by dynamic path planning. Several AGVs' paths are planned simultaneously in real time and path optimization is achieved with a heuristic algorithm in dynamic path planning. It is proved with simulation that the strategy improves flexibility of AGV systems and is a universal scheduling strategy for different maps.