The patellofemoral joint (PFJ) plays an important role in the extensor mechanism of the knee. Several types of PFJ disorders are commonly found in about 25% of the people. It is believed that patellofemoral (PF) disorders, e.g. excessive lateral pressure syndrome and patellar maltracking, may be associated with articular cartilage contact pressure elevation, which accelerates degenerative joint disease and causes anterior knee pain. To reduce the pressures, a number of anatomical interventions have been applied to correct contact mechanics and patellar tracking. However, the rate of successful surgery is not high because the anatomical complexity of the joint itself and complex symptoms make diagnosis difficult. For this reason, various computational modelling techniques have been developed to assist in diagnosis and prognosis of PF disorders. This research aims to develop a finite element (FE) modelling method and study the feasibility of its clinical applications. The modelling methods may assist in the diagnostic and treatment planning processes. The research was divided into five phases: 1) development of an FE modelling method to analyse PFJ models 2) model validation using in vitro experimental data 3) development of subject-specific input estimation method from routine diagnosis protocols 4) model sensitivity analysis and 5) clinical applications. The FE results included joint contact force, contact pressure, subchondral bone stress and patellar kinematics. The validation and sensitivity analysis showed that the FE modelling method could adequately analyse PFJ biomechanics. Approval for a clinical study was obtained from the National Health Service (NHS) Research Ethics Committee, and groups of control subjects, anterior knee pain (AKP) patients and those with trochlear dysplasia and trochleoplasty were recruited. The modelling method was applied to analyse their knees and predict their non-operative and operative treatment outcomes. The study showed that the biomechanical responses of the PFJ and the treatment evaluations were variable. In particular, it was found that AKP was associated with significant elevation of contact pressure; thus confirming the usefulness of the FE modelling method as a powerful diagnostic and surgical planning tool for subject-specific PFJ treatment.