Abstract: In order to address the operational challenges brought by large-scale wind power grid integration, this paper systematically analyzes the impact mechanism of its output randomness and intermittent fluctuations on grid frequency, voltage, and power flow, and explores key technologies for power prediction. On this basis, focusing on improving system flexibility and resource coordination efficiency, a multi-source collaborative optimization scheduling system covering traditional power sources, energy storage, demand side response, and virtual power plants was constructed. Advanced methods such as multi scenario modeling and multi-scale rolling scheduling based on stochastic optimization were deeply studied, aiming to provide reference for the safe and economic operation of high proportion renewable energy power systems and promote the development of intelligent and refined power grid scheduling.