Microscopic traffic models are developed to simulate traffic phenomena. However, current models are usually based on the fact that vehicles travel in the middle of a single lane and make a lateral movement only when the driver wants to change lane. They are incapable of describing driving behavior in a complex traffic environment with a lot of lateral separations which may cause traffic flow to be unstable on a single lane. In order to solve the problem, in this paper we first build up the equations of motion based on target steering angle (TSA) and then establish a series of rules for determining TSA by considering the lateral separation characteristics between the follower and the leader. Finally a new vehicular behavior model is developed. The properties of the model are investigated by simulating two scenarios. It is found that the proposed model is capable of describing complex phenomena which cannot be described by existing models and explaining the perturbation caused by bus near bus stations. The results also imply that this model with taking the lateral separation characteristic into account greatly enhances the effectiveness of vehicular behavior.