A new model describing human behavior in the control loop consisting of vehicle, driver, and environment is described. This driver model was developed to enable realistic closed-loop maneuvering in handling dynamics simulation - incorporating specifically human behavioral patterns. It is also suitable for deriving optimal-performance driving, e. g. in motorsport applications. Experience and preferences of a driver can thereby specifically be varied. The driver model's performance is demonstrated using a fast evasive maneuver (ISO double lane change). The calculated behavior, and in particular the stability margin at highly dynamic excitation, correspond well with human behavior. When developed further, the driver model is a very good basis to give a realistic and reliable prediction of the human controller's action, especially its adaptation to different vehicles.