Virtual power plants are an effective approach for achieving efficient management of high-penetration distributed energy resources and various flexible resources on the source side, grid side, load side, and storage side. To further exploit the internal flexibility of virtual power plants while fully utilizing the complementary characteristics of resources among multiple entities, this paper proposes a hybrid game based operational optimization method for multiple virtual power plants considering dynamic pricing incentives for flexible resources. The distribution system operator acts as the global coordinator, promotes energy sharing among virtual power plants through price signals, and represents virtual power plants in the electricity market and ancillary service market. First, a dynamic pricing incentive model driven by multiple factors is constructed for electric vehicles and shiftable loads. Second, a bi-level optimization model based on a Stackelberg and non-cooperative hybrid game is established. In the upper level, the distribution system operator determines electricity buying and selling prices. In the lower level, each virtual power plant formulates incentive prices and internal scheduling strategies while considering the uncertainty of renewable energy output. Finally, an improved centered collision optimizer integrating opposition-based learning and differential evolution is designed and combined with the alternating direction method of multipliers to solve the model, thereby achieving efficient computation and privacy protection. Simulation results verify the effectiveness of the proposed method. The results show that the proposed method can significantly promote energy interaction among virtual power plants, improve the utilization efficiency of flexible resources, and reduce the operating costs of virtual power plants.