Abstract: To enhance the accuracy of the virtual power plant scheduling plan and fully exploit the adjustment potential of flexible resources in the virtual power plant, in this paper we propose a day-ahead and intraday multi-time scale optimization scheduling method that makes both day-ahead and intraday forecast of wind and solar output as well as load. Additionally, we utilize energy storage batteries to smooth the fluctuations of intraday wind and solar output, employ solar-thermal power station to decouple the thermoelectric units with the aim of mitigating their minimum output, and consider demand response resources with varying response speeds. A multi-time scale thermoelectric joint optimization scheduling model for virtual power plants containing solar-thermal power stations is established, and the adaptive genetic algorithm is employed to solve it. The simulation results of the example show that, compared to traditional day-ahead scheduling, the multi-time scale optimal dispatching enables a more refined dispatching plan, thereby further promoting wind and solar consumption, reducing the output deviation of the virtual power plant, and improving the economy of the virtual power plant.