Abstract: Aiming at the deadlock problem of artificial potential field (APF) and the lost problem of feasible workspace in grid map, a hierarchical path planning strategy based on gridgeometric map was proposed. By adopting improved A* algorithm in upper layer to search for a basic route in highclass granularity grid map, both the deadlock problem of APF and backtracking problem in traditional A* algorithm can be overcome. To reduce buffeting problem, inertial artificial potential field (IAPF) was introduced to smooth the basic route in bottom layer. Simulation results show that the proposed strategy can effectively decrease computational consumption, and the searched route is obviously better than the one according to pure grid method or IAPF method.