Microgrid, which can combine different forms of renewable energy generation, effectively retards the expansion needs of the grid. Power capacity design is the basis of microgrid planning and construction, which can also provide reference for the economy, environmental friendliness and power supply reliability of microgrid construction in the planning and design period. The results of the optimal capacity design of microgrid largely depend on the choice of the optimization model, different optimization models will result in different capacity allocation results. Therefore, this paper focuses on the influence of different optimization models on the results of capacity allocation, and analyses the differences and causes of photovoltaic, hydropower, diesel power and battery based standalone microgrid capacity allocation results. Optimization models with different optimization objectives are established with comprehensive consideration of economy, environmental friendliness and power supply reliability. An actual microgrid project is used as an example to solve the models by adaptive weighted particle swarm algorithm, and optimum solutions of capacity allocation are obtained in different models. The differences of different optimal capacity design results are analyzed, as well as the reason, and the correctness and reliability of the proposed model are verified.