Abstract: In order to solve the problems of low gangue picking rate and low utilization rate caused by no strategic planning in the process of sorting random gangue on belt conveyor by coal and gangue sorting robot, the mathematical model of multi-dynamic and multi-manipulator cooperative sorting strategy under random gangue flow is constructed according to modern sorting theory, robot technology and multi-objective planning method.According to the representation of D-H robot, combining with geometric parameters of multi-manipulator module and belt conveyor and camera parameters of visual identification module, the coordinate transformation relationship between the gangue image and the world coordinate system and the end effector relative to the world coordinate system is established respectively.Considering the static time of multi-manipulator in the process of gangue flow movement, in this paper, the pre-feedback, self-feedback and cooperative feedback mechanism of the double rectangular coordinate robot are proposed to realize the increasing trend of the residual value of the sorting area in the collaborative sorting process of the multi-mechanical arm.According to the performance parameters of manipulator and belt conveyor, the mathematical model of multi-manipulator cooperative sorting strategy is constructed to realize the accurate and efficient sorting of multi-dynamic target gangue.Based on the strategy solution results, the gangue model is changed, and the influencing factors of multi-manipulator cooperative sorting multi-dynamic target gangue sorting rate are summarized.According to the simulation experiment result, the relevant parameters of the experimental platform of coal and gangue sorting robot are modified, Through the comparative analysis of the results of the capture experiments of multiple groups of gangue flow, in the belt conveyor, the stone sorting rate meets the application requirements, and the total feedback stroke of the mechanical arm is inversely proportional to the number of the sorting.