Acoustic emission (AE) signals initiated by mechanical faults or damages is composed of two types of signals: high frequency burst impulse signal and long period quasi-stationary noise signal. Wavelet scalogram has a particular time-frequency localization, which helps it to be well used for describing the time-frequency characteristics of AE signals. By analyzing the characteristics and feature extraction of typical AE signals, the paper applies wavelet scalogram for fault diagnosis based on AE technique, and presents the wavelet scalogram analysis method of AE signal for the first time. By theoretical analysis and simulation, the wavelet basis function and parameter related to the function are defined. So the limitation that best time resolution and frequency resolution of wavelet scalogram cannot get at the same time is overcome effectively. When applying wavelet scalogram for fault diagnosis of rolling bearings based on AE techniques, the results are quite visualized, clear and accurate. Both simulations and experimental research prove that wavelet scalogram can be used for condition monitoring and fault diagnosis based on AE detection well.