There are uncertainties in time and space for electric vehicle travel. Research on load prediction of EVs considering the spatial and temporal distribution is the basis of future research on the interaction between EVs and the power grid and an orderly charging control strategy. Taking electric private cars as the research object, this paper proposes a method for load prediction of electric vehicles based on an OD matrix and considering spatial and temporal distribution. First, a probability model of basic parameters of EV charging load is established according to the influencing factors of EV user travel habits, charging behavior characteristics and charging mode. Secondly, a topology model is established according to the actual road network. The OD matrix is combined with a Floyd algorithm to simulate the shortest distance travel trajectory of electric vehicles. Then, considering the continuous change of the battery charge state, a charging load prediction model of EV is established based on the Monte Carlo method. Finally, the method is used to calculate the spatial and temporal distribution of EV charging load in a city including residential, business and working areas. Results show that the electric vehicle charging load is obviously regional, with residential area charging load mainly concentrated in the day after 19:00 to 05:00, and commercial and workspace concentrated in the day, from 11：00—17：00. The electric vehicle charging load increases the peak load of the distribution network, and total load affects the safe operation of the distribution network. The results provide base data for strategic research of the orderly charging of electric vehicles and the locating and sizing of charging stations. This work is supported by the National Natural Science Foundation of China (No. 51307152).