In order to realize the tracking control of arbitrary path in the actual tactical exercise of robot tactical target, a strategy of teaching path sampling and tracking control is proposed. The path point sampling strategy is established based on the 2 teaching methods, in which the path point sampling strategy based on the driving distance and turning radius is established for the online teaching of manual remote operation of robot tactical target walking, and the teaching path is drawn by editing and defining a series of discrete path points on the map file through the host computer for offline teaching. In the path reproduction stage, the taught path stored before is read, and the path tracking of the robot is controlled by using a node tracking switching control strategy, a speed segmentation control strategy and a direction control strategy based on a heading and offset combined error. The algorithm is verified on the physical prototype, and the results show that the robot tactical target can track any complex teaching path well, in which the path tracking deviation is less than 0.3 m, and the point-to-point tracking speed deviation of each path is less than 5%.