A community integrated energy system integrates various types of energy source to provide a reliable energy supply. However, the off-design condition characteristics of the equipment pose serious challenges to the rationality of capacity configuration and the accuracy of scheduling schemes. For this reason, a two-layer optimization method for capacity and scheduling of an integrated energy system considering off-design condition characteristics is proposed. First, an improved kernel density estimation method is used to construct more accurate energy supply and demand scenarios. Secondly, based on an extreme gradient boosting algorithm, an equipment efficiency correction model is established to characterize the off-design condition features of the equipment. And considering the impact of temperature on equipment life, aging costs are incorporated into the objective function for optimization. Then, by proposing an optimal load rate tracking strategy, this paper decouples the relationships among output, efficiency, and load rate during the scheduling process. By actively adjusting the output ratio of energy conversion devices, a flexible energy supply system is achieved. Finally, simulation results show that the proposed method takes into account the off-design condition characteristics of the equipment, not only improving the accuracy of the equipment model but also enhancing the scientific nature of the configuration plan and the economic aspect of the scheduling scheme.