Abstract: The scheduling system of modern intelligent coal mine underground transportation locomotives is the main link to realize coal mine safety and efficient production. It has the characteristics of multi-terminal processor, information network transmission and information intelligent decision-making. It is helpful to realize a safe and efficient automatic scheduling by studying the information interaction process of the terminal processor of the scheduling system in the scheduling process, optimizing the functions of the terminal processors of different network nodes, improving the reliability of information transmission in the scheduling system, reducing the delay time of information transmission, and constructing a rapidity and reliability model of locomotive route switching scheduling in the scheduling system. Based on the network topology of KJ654 system dispatching information transmission, the cyber-physical system theory and unified modeling language are used to create the cyber-physical interaction model of multi-terminal controller coordinated control of locomotive dispatching system. Firstly, the distributed structure of the interaction between the terminal information layer and the physical layer of the locomotive dispatching system is established by using the theory of cyber physical system, and the composition and function of the information layer and the physical layer are planned. Based on the automatic safe and fast switching function of the locomotive route, the tasks completed by each terminal processor and the need to exchange transmission information are studied. From the perspective of controller task execution, the information interaction sequence diagram between controllers in the process of locomotive route switching is established, and the task priority and execution time requirement in locomotive dispatching are merged. The cyber physical interaction model of controller cooperative control between different regions is constructed. Finally, the information interaction scale of locomotives in different areas and the information interaction time of different route switching are obtained by experiment and simulation. The results show that the KJ654 terminal processor program written based on the physical information interaction model runs in the network architecture invariant scheduling system. Compared with the original system, the response time of route switching request is reduced, the locomotive running time is significantly reduced, the reliability of processor operation is improved, and the regional distributed control decision-making ability is enhanced. This modeling method clearly describes the information interaction process between controllers, ensures the logical rigor of the model, optimizes the task and information transmission content of the terminal controller of the scheduling system, and provides a method for the multi-processor coordinated control of locomotive scheduling.