Abstract: For knowledge express and learning, A novel conception of graphical topological network knowledge express ba-sed on neural network and fuzz logic rules and relativeknowled-ge learning method are put forward. It can express knowledge neatly, form knowledge warehouse conveniently and straight, make for learn from intelligent system. Based on the reciprocity of expert system, fuzz logic and neural network etc. in common use in intelligent reasoning, and combined general knowledge pool of graphical topological network knowledge express, the integrated reasoning model are brought forward, it can form better reasoning function, apply and deal with intelligent system availably. This achieved in practice of device status monitoring and diagnosis.

Key words: Expert system, Fuzz logic, Intelligent system, Monitoring and diagnosis, Neural network