The invention relates to a multi-stage rolling optimization planning method for a distributed energy system. The multi-stage rolling optimization planning method comprises the following steps: S1, building an internal and external thermoelectricity decoupling simulation evaluation model, and building an iterative optimization operation flow of the simulation evaluation model; S101, establishing anelectric and thermal balance equation; S102, according to an outer-layer bus balance equation and a scheduling strategy, obtaining an outer-layer bus balance equation; S103, inputting the bus exchange electric/thermal power determined by the outer layer into the electric and thermal system model of the inner layer in the step S101, and judging whether the change of the internal network loss valueof each bus before and after calculation is smaller than an allowable error range or not; S105, outputting the system state quantity as a simulation result; S2, establishing an intelligent optimization model; and S3, carrying out rolling optimization solution on the simulation evaluation model and the intelligent optimization model. According to the multi-stage rolling optimization planning method, optimal configuration of the distributed energy system is realized, and the precision and expandability of location and volume determination of the distributed energy system are improved.