Digital human models (DHM) are widely used in automotive industry to simulate the driver in a very early stages of production, where no physical prototypes of the car exist. In case of crash simulation, detailed finite element models of the human body are used to simulate the highly dynamic impact and the resulting injuries in the human body. Models with multibody kinematics are widely used, when the reachability and the ergonomic assessment of the driver is investigated. These kind of models are only used in quasit static scenarios where the car is standing or driving with constant velocity. In dynamic driving scenarios like cornering, sudden breaking or pre-crash scenarios, both types of models are not applicable. The FEM models are much too time consuming, because in contradiction to crash simulation the simulated time span is bigger. Also these models are difficult to control. The kinematic models are not able to take into account dynamic loads and contact forces. Also the motion generation is difficult, because the usually base on forward or inverse kinematics. In this work we will present an approach, how to enhance a multibody based DHM to generate human like motion for dynamic driving maneuvers. Therefore, the human is modeled as a multibody system, where the limbs are the rigid bodies, which are connected via joints. Hill muscles are used to actuated the multibody system. These are digital versions of the real muscles in the human body. To generate the dynamic human motion an optimal control algorithm is developed, which is able to handle opening and close contacts. These enables to simulate the dynamic interaction of the DHM with the car interior like seat, pedals or steering wheel. In this approach only some basic boundary conditions must be described, like at the start the human is sitting at a certain positon with two hands on the steering wheel and the trajectory of the car. With a certain objective function, the optimal control approach than generates the desired control (muscle actuation) and the human motion.