Single train model of operation energy consumption and power supply system was analyzed firstly, and on the basis of this, a multi-train operation energy consumption model was established. In order to make full use of the regenerative braking energy, a method was put forward to set up the train operation schedule which could implement fine tune of departure interval, the stopping time and the running time. A multiple population genetic algorithm was designed to optimize the train operation schedule with the minimum energy consumption and the maximum utilization of the regenerative braking energy as the goal. The elite reservation strategy and migration strategy were adopted to improve the stability and speed of the optimization process in the algorithm. Simulation results showed that this method could increase the utilization rate of regenerative braking energy from 14.8% to 80.9%, and the ratio of regenerative braking energy utilization and operation energy consumption could be increased from 0.74% to 4.2%.