The unanticipated fault diagnosis for complex system is a difficult problem. A general process model was proposed. The model adopted a four-layer structure and included four models, i.e. an anticipated fault detection model, an anticipated fault identification model, an unanticipated fault detection model and an unanticipated fault identification model. Several key problems and the corresponding algorithms in each layer were analyzed, including the establishment of the fault detection statistic, the extraction of the fault character, the design of the fault isolation criterion and the construction of the fault contribution rate. The general process model regularized the diagnosis process of unanticipated fault for complex system and defined the datadriven fault diagnosis framework. The effectiveness of the proposed general process model was validated by a satellite control system. 