A novel AE(acoustic emission) source localization algorithm by combining wavelet transform and threshold analysis together is presented aiming at the features of AE signals’ multimodes and low SNR(signal noise ratio).The signal spectral analysis of the collected acoustic emission signals is carried out to extract the wavelet transforms amplitude which is one of the main frequency components,and the timedomain waveform containing multiple peaks is gotten.80% of the maximum amplitude of the wavelet transforms is set as the threshold to eliminate the influence of the multiple peaks.Thus,through getting the time difference of the same wave mode under the same frequency of the AE signal arriving different sensors the time difference of the AE signal arriving different sensors is achieved.After measuring the group velocity of the wave mode,AE source localization is realized. Experiments of the location of AE source between two points on a marble slab are carried out.. Results show that the method is an effective method for locating the AE source and can improve the accuracy of AE source localization.