With the development of economy, new energy power generation based on wind power, photovoltaic power and solar heat has developed rapidly. However, due to the instability of multi-energy power generation, the phenomenon of wind and light abandonment is serious. Therefore, based on the regulation characteristics of the photovoltaic power station, this paper puts forward the joint power generation scheduling strategy of wind power, photovoltaic power and photovoltaic power. With the good scheduling and controllability of solar thermal power generation, the randomness and uncertainty of wind and scenery output are reduced, and the “peak cutting and valley filling” is realized. Aiming at the lowest comprehensive cost and the lowest power fluctuation, the output of each unit is calculated. The improved multi-objective particle swarm optimization algorithm was used to solve the model. Finally, the feasibility and effectiveness of the proposed method are verified by taking the actual data of a northwest region as an example.