In industrial practice parallel kinematic structures start to replace serial ones, in case of time-sensitive, highly dynamical applications or processes requiring a high structural rigidity of the robots. But the high complexity of their dynamic models is still the biggest impediment for the realization of model based trajectory tracking control strategies.In this thesis a holistic concept for the model-based trajectory tracking control of a special class of parallel robots - containing functional or structural redundancies (complex parallel robots) - is introduced. Since for these robots - due to the high complexity of their dynamic models - sophisticated model-based strategies for trajectory tracking control cannot be directly applied, a new concept is necessary. It mainly bases upon the derivation of real-time implementable compact dynamic models for the studied class of robots. But compact dynamic models are only a first step on the way to a model-based trajectory tracking control with high path accuracy and a low traveling time. Additionally, appropriate designed trajectories are required. For planning them a new strategy has been developed and successfully applied.The holistic approach covering also aspects like parameter identification and velocity estimation has been successfully demonstrated at a planar parallel robot including experimental results.