In this paper, we tried to represent the driver model which would be able to describe comparison between beginner driver and expert driver. Because it is very important to develop effective driver assist system. We focus our attention on steering maneuver. We assert that driver use various function to recognize vehicle maneuver and environment around vehicle for drive. We assume that driver recognize them by visual modal, somatic modal and reaction modal. So we suggest "Sensorimotor by Multi-modal driver model". We simulate the driver skills of beginner, normal and expert by that model, and what's more we try to focus the analysis of expert driver skill.