Abstract: A comprehensive intelligent fault diagnostic method with multi parameters is proposed. This method combines process and vibration parameters, adopts hierarchical and categorized diagnostic strategy and integrates rules based model, cases based model, fuzzy logic and neural networks. The necessity to integrate multiple fault diagnosis approaches is analyzed. Furthermore, the knowledge representation, intelligent inference and knowledge learning of this method are presented. This method has been applied to a condition monitoring and fault diagnosis system. The industry application shows this method is useful and effective.