According to the developed trend for the combination of microwave and laser links in the future data relay satellite system, the resources scheduling problem in data relay satellite system based on microwave and laser links is studied. A multi-objective constrained programming model for microwave and laser links data relay satellites scheduling problem (ML-DRSSP) is presented under the constraint condition of visible time window, mission priority and power consumption of the terminal. By designing the self-adaptive cross, mutation operator and selection operator based on elitism selection, a modified non-domination sorting genetic algorithm Ⅱis improved and the model is solved. A scene simulation is set up on the conditions of a data relay satellite, four user satellites, three resource terminals and scale of different tasks. The simulation results show that the algorithm, by keeping the solution diversity and making non-inferior solution set close as possible to the Pareto optimal solution set, can solve ML-DRSSP with multi-tasking and multi-type antennas, effectively.