Integrated energy systems have a significant role and non-negligible importance in modern energy systems. In order to improve the operating economy of the integrated energy system and at the same time to cope with the challenges brought by multiple uncertainties within the system, an economic scheduling model of the integrated energy system is proposed and solved by deep reinforcement learning method. Firstly, the optimal scheduling problem of the integrated energy system is mathematically modeled; secondly, it is converted into the expression form of Markov decision process, and the state, action space and reward function are established; after that, the DQN algorithm is improved to increase the training speed and stability of the algorithm, and the optimal scheduling problem is solved with the goal of economy. Finally, the effectiveness and superiority of the proposed method in optimized scheduling of integrated energy systems are verified through arithmetic simulation and comparison with traditional algorithms.