Abstract: With the exhaustion of fossil energy, distributed energy system has become a research hotspot. On the basis of isolated community load modeling, this paper builds a distributed energy system composed by photovoltaic array, wind generator, diesel generator and battery power supply, and proposes a multi-objective nonlinear optimization model that takes into account net present value cost, environmental penalty and renewable energy utilization hours simultaneously. The optimization variable was wind generator type, wind generator number, photovoltaic array capacity, diesel generator type, diesel generator number, battery type and battery number. A non-dominant sorting genetic algorithm with elite strategy was used to solve the optimization model. The relationship between distributed energy sub-system is obtained by analyzing component life sensitivity. Moreover, taking a residential community micro-grid as an example was studied and compared with calculation results of HOMER. Finally, the optimal equipment capacity of the residential community micro-grid was obtained. The results showed that the proposed optimization design method was reasonable and feasible, providing a case for engineering design.