Abstract: In order to diagnose the low speed fault rolling bearing and avoid extremely weak signals, it is suggested to use traditional vibration examination method. AE inspection experiments to the different fault pattern of rolling bearings were carried under the laboratory condition. Then the AE signals obtained from rolling bearing were analyzed using parameters and wavelet. Eventually fault patterns of rolling bearing were synthetically judged using hits and neural networks. The validity and the accuracy of low speed bearings fault diagnosis can be enhanced.