Abstract: In complex flexible manufacturing systems, it is important to conduct reasonable autonomous mobile robot (AMR) scheduling. In this study, the path planning and task assignment of AMR scheduling are studied to minimize the time for AMR to complete tasks. A mathematical model is established with the minimum AMR task completion time as objective function, and the map is modeled based on topology method. A greedy algorithm is used to assign tasks for orders with a constant length distribution for the interval time, and the algorithm calculation is reduced by classifying the AMR working state. Global path planning is based on the Dijkstra algorithm to search for the global shortest path of the AMR, and local obstacle avoidance path planning is performed through the AMR's laser radar. Finally, the effectiveness is verified through scheduling simulation experiments on the openTCS platform.