Musculoskeletal disorders (MSDs) are chronic pathologies affecting a large population across the world with significant social impact. Among these pathologies, hip osteoarthritis (OA) is characterized by the breakdown or the degeneration of the articular layers of joints causing symptoms such as pain and limitation of movement. Different studies have shown that hip OA is a multi-factor disease. Nevertheless, hip OA is considered idiopathic and some causes are still not completely known because the development of this pathology generally evolves over a long time period. Among the different hypotheses that were proposed to explain the mechanisms of hip OA degeneration, various studies focused on the role of hip kinetics (forces/loading) and kinematics (movements). For example, athletes practicing sports that incur stress on hip joints or ballet dancers performing excessive hip motions are considered as a population at a higher risk for the development of early hip OA. In this thesis, we investigated the possible links between OA and hip mechanics based on a study with professional ballet dancers. We focused on the design of an efficient protocol to analyze the hip joint mechanics during dancing movements. Given the numerous differences that exist between individuals, the use of subject-specific models is of paramount importance to individualize the diagnosis. To this end, our research innovates in the derivation of a methodology based on non-invasive methods and subject-specific data to yield an effective biomechanical modeling of hip joints. Anatomical, kinematical and kinetical models resulting from techniques, which have their origin in computer graphics, were combined to achieve accurate physically-based simulations. In addition to the investigation of different scenarios, the simulation allows to assess and to visualize the articular deformation of the hip joint during movements. Our investigation established a strong correlation between simulation results and clinical observations which included clinical and radiological findings. We demonstrated the potential risks of excessive movements which contribute to the development of early hip OA. We conducted various experiments to support this investigation by focusing on various performance indicators such as the robustness and accuracy of the simulations. The proposed methodology is not restricted to the hip as it provides general guidelines to build efficient computational modeling approaches to investigate and analyze MSDs. In addition, we exemplified the flexibility of our simulation framework to support different types of applications not necessarily related to the biomedical domain (e.g., realistic animation of the human body by simulating the skin deformation). In the long-term, we believe that our work could serve as a template for studying the biomechanics of articulations and could also be adapted to the development of new computer-aided systems to support practitioners in clinical applications such as diagnosis, predictive medicine, surgical planning, postoperative assessment and rehabilitation planning.