Abstract: The scheduling optimization of integrated energy distribution system has many problems, such as the high uncertainty of both sides of the source and load, the complexity of multi energy flow coupling and the lack of real-time response ability. Therefore, a scheduling technology of integrated energy distribution system based on improved generation countermeasure network is proposed. In this paper, an integrated energy distribution system model including the coupling of electricity, heat and gas is constructed, and the gradient penalty mechanism is introduced into the traditional generation countermeasure network, which significantly improves the data generation quality and generalization ability of the model in complex energy scenarios. The scene samples generated by the improved generation countermeasure network are integrated with the deep reinforcement learning framework, and a multi-objective scheduling optimization model is constructed. The experimental results show that the proposed technology significantly improves the renewable energy consumption rate and reduces the system operation cost, which provides a new idea for the intelligent scheduling of integrated energy distribution system under the condition of high proportion of renewable energy access.