In order to configure accurately the capacity of grid-connected distributed generation system, a method of distributed generation system capacity optimization and configuration based on power generation efficiency attenuation, time-of-use electricity price and load increasing year by year which aims at the grid-connected wind-photovoltaic hybrid power generation system is given. Using the wind speed and sunshine radiation data of NASA's meteorological sites, annual output curve of unit capacity wind-photovoltaic power supply is obtained, the distributed generation planning model is established aiming at fixed load and increasing load in the scenario of single electricity price, time-of-use electricity price and elimination on the spot. In the case analysis, genetic algorithm is used to solve the equation, the method’s validity and correctness are verified. The model is more close to the actual engineering application, and it is easy to expand to grid-connected power generation system that includes a variety of distributed power, it is helpful for the application and promotion of grid-connected distributed generation system.