Abstract: Optimizing and adjusting train diagram routing is of great significance for saving operation cost, improving services quality and operating efficiency. A PSO (particle swarm optimization) algorithm based on SA (simulated annealing) algorithm is proposed to optimize and to adjust train diagram. Taking the minimization of passenger waiting time as the objective function, an optimization model of train diagram routing is constructed. A PSO algorithm based on SA algorithm is designed to get the global optimal solution of variables. Taking Nanning Rail Transit Line 1 as an example, the departure time, departure interval and other elements of the train timetable are optimized, and the compatibility between the optimized train capacity and the traffic volume is analyzed. It is found through practical case that the proposed model and algorithm can tackle the issues of train capacity and traffic volume well, improving the passenger ride comfort.