Based on wavelet analysis of the emitters, a new fingerprint feature ext raction method for emitter identification based on wavelet entropy is proposed. Firstly, the signal power spectra are calculated. Secondly, the wavelet coeffici e nts are extracted by continue wavelet transform. Finally, the wavelet entro py is extracted as a feature vector. Using neural network classifier, the compar ative experiments with traditional square integral bispectrum are carried out ba sed on twenty interphones. The experimental results show that the method can achieve individual classification by transferring the signal time frequenc y characteristics to the feature vectors through the entropy of the wavelet coeff icients. Besides, the proposed method is insensitive to noise, and the system re cognition rate is above 95% and more than 80% with SNRs of 20 dB and 5 dB,respectively.