Abstract: Parameter calibration of microscopic traffic simulation model is the foundation and prerequisites to ensure the effectiveness of simulation models .This paper proposes a parameter calibration method of microscopic traffic simula‐tion model focusing on the different driving behaviors on the main roads and side roads .The errors of the average delays on road entrances are selected as an evaluation indicator .Genetic algorithm is used to search the optimal parameters .The parameters of driving behaviors on the main roads and side roads in Vissim are calibrated .The driving behavior parame‐ters were calibrated using the proposed method in the model of the intersection of Chaoyang road and Gaobeidian North Road as well as the intersection of Nanyuan road and Jiujingzhuang road .The results show that the errors of the average delays were 2 .1 s on main road and 2 .5 s on side road by using the proposed calibration method .While using the unified method ,the errors of the average delay were 7 .1 s on main road and 7 .5 s on side road is 7 .5 s .The study concludes that the proposed method is effective and improves the efficiency and accuracy of the simulation .