Despite the rapid development of electric vehicles (EVs), the limited range and accessibility of charging infrastructure are still obstacles to the development of EVs. At present, the Chinese government has made great efforts to set up charging stations and the most popular construction mode of charging stations in China is to divide a certain proportion of parking spaces in the parking lot to build charge piles. The proportion for every parking lot is the same and recommended by the government, which is a one-size-fits-all construction strategy. Nevertheless, with the growth of EVs, some challenges arise such as how to improve the accessibility of charging facilities in hot parking lots. The current work addresses this challenge by correctly forecasting the charging demand of EVs based on simulation. In particular, in this paper, based on agent simulation technology, electric vehicle agent, fuel vehicle agent and parking lot agent are established to predict the charging demand and the Wulin business district of Hangzhou City is used as an example to predict the charging demand of nine public parking lots. Simulation results show that under the current number of EVs, this one-size-fits-all setting strategy can meet the demand with high accessibility. However, with the increasing of EVs, the spatial distribution of charging demand will be more uneven and a decline in system overall benefits will appear, considering both EVs charging and fuel vehicles parking. Given all this, an optimization method based on GA (Genetic Algorithm) is put forward and an optimized setting strategy is proposed.