It is significant to develop Web system of intelligent fault diagnosis for steam turbine based on artificial neural networks and Internet technology as the rapid progresses of Internet, modern communication, fault diagnosis and neural networks. Vibration fault diagnosis of steam turbine is discussed. With combining improved neural network and object-oriented paradigm, a neural network system of intelligent vibration fault diagnosis for steam turbine is constructed using four symptom groups and five BP networks. Automatic obtaining method of vibration spectrum symptom of steam turbine is presented. Two fault diagnosis examples based on spectrum symptom are given. A Web system of intelligent fault diagnosis of steam turbine using browser/server mode is developed. The system configuration, function modules and Web design method are described.