Abstract: In accordanc with multiple stakeholders and uneven income distribution in the park-integrated energy system, a two-layer energy management model based on a hybrid game theory is proposed. Firstly, the operational framework of the park-integrated energy system is established, and the interests of the upper-layer microgrid operators and the lower-level user aggregators are analyzed. Secondly, to maximize the interests of all stakeholders in the park, a hybrid game model is developed, allowing for multi-user and multi-party participation of microgrid operators. The operator formulates the price of energy and sell to users through the Stackelberg game, while the user aggregator distributes the benefits based on the Nash-Harsanyi theory after receiving the price. In light of the substantial initial investment in energy storage equipment, the potential for using electric vehicles as a cluster dispatchable resource is thoroughly investigated. Historical data from electric vehicles are analyzed by using the CNN-BiLSTM method, which combines convolutional neural networks and Bi-directional long short-term memory, to mitigate uncertainty. The strategy has been developed to utilize the shared energy storage capabilities of electric vehicles as energy storage devices. Finally, an integrated energy system of a city is taken for example to perform analyses. The results show that the established model can effectively decrease carbon emissions and create a win-win situation for operators and multiple users.