An effective energy management system is a critical technology for the secure and economical operation of microgrids. This paper comprehensively considers the output of renewable energy sources and the flexibility of demand-side loads, and proposes a microgrid energy management method based on a state potential game (SPG). First, user loads are categorized into basic loads and various flexible loads while fully considering the flexible resources on the user side, and models for microgrid entities are developed. Next, an SPG model is established by treating each component unit within the microgrid as an autonomous decision-making entity. Finally, a genetic algorithm is embedded in the outer layer of the SPG to implement periodic constraints on shiftable and transferable loads. Simulation results show that, while ensuring the autonomy of each entity, the proposed method significantly improves the integration of renewable energy and the economic efficiency of electricity consumption for users. Additionally, it enables the virtual charging and discharging of flexible loads. Compared to pure potential games and non-game methods, the proposed method better reflects the equal status of all entities while simultaneously minimizing communication requirements among them to enhance privacy protection.