Abstract This paper presents an intelligent methodology for diagnosing incipient faults in rotating machinery. In this fault diagnosis system, wavelet transform techniques are used in combination with a function approximation model to extract fault features. Wavelet neural networks are also constructed. The main contributions of this paper are as follows: First, a wavelet theory based on a nonlinear adaptive algorithm is developed for an excitation function approximation of neural networks. Preprocessing of a single fault signal is required to perform diagnosis using an intelligent system. Second, a neural network classifier for identifying the faults is developed. The system is scalable to different rotating machinery and has been successfully demonstrated with a turbine generator unit.