Real-time capable and accurate simulation models of vehicle dynamics for the further development of control and diagnostics in motor vehicles can achieve nearly optimal simulation results when the models are automatically adapted to changes in vehicle parameters or environmental conditions. The combination of a physical vehicle model with parameter estimation methods and neural networks enables a significant improvement in simulation quality compared to conventional vehicle simulation models. For practical implementation, the advantages of the adaptive model concept are demonstrated through several application examples, and the real-time implementation of the adaptive model in the test vehicle is also discussed. The findings relate to vehicle dynamics, model development and simulation technology, as well as identification methods and model-based development.