This paper describes a novel wavelet-based approach to the abrupt fault detection and diagnosis of sensors. By the use of wavelet transforms that accurately localize the characteristics of a signal both in the time and frequency domains, the occurring instants of abnormal status of a sensor in the output signal can be identified by the multiscale representation of the signal. Once the instants are detected, the distribution differences of the signal energy on all decomposed wavelet scales of the signal before and after the instants are used to claim and classify the sensor faults. Synthetic data simulated by means of a computer using real-word data from a general-purpose pressure sensor have verified the effectiveness of the proposed method.