Abstract: The assistant decision system for train regulation is a key link of train operation adjustment auto-control system. Based on analysis and design of the workflow and components of the system, a mixed 0-1 linear programming model for train operation adjustment is built. Against the disadvantages of the traditional processing method, a finite state policy used for the process control of regulation is taken to improve the performance of process. This scheme is not only in agreement with the basic idea of structured system design, but also easy to realize in software and practicable