In order to optimize the service process of towing tractors, the problem of tractors dispatching optimization is studied through theoretical calculation and case verification. It combined with Airport Collaborative Decision Making (A-CDM) system, mathematical model and multi-objective genetic algorithm. Firstly, the plan time of Push-Back of each aircraft was obtained by the A-CDM system. And the time-consuming of operations and tractors travelling was designed based on the historical operation data. Constraints were set based on the conflicts during the operation of Push-Back. The mathematical model in a single area within a certain period of time was established. Objective function included three parts: the smallest total cost caused by the Push-Back service, the smallest number of tractors participating in the operation and the balanced tasks distribution. Then, the multi-objective genetic algorithm was designed and verified by an example. The results show that the proposed method has been significantly optimized compared with the traditional manual scheduling model. The total cost of operations is reduced by 31.67%, the variance of the missions is reduced to zero, and the formulation time of scheme is shortened by 71.15%. Finally, a visualized program of scheduling tractors is designed, which provides decision-making reference for airport operators.