Nuclear quadrupole qesonance(NQR) methods can be used to detect hidden explosive,with advantages such as high accuracy,low false alarm rate and non-radioactive contamination.However,the practical use of the NQR methods for explosive detection are limited by low signal-to-noise ratio.In this work,a signal process method for NQR signals is presented,which is based on wavelet transform and specially designed for the detection of pentaerythritol tetranitrate(PETN) with NQR.It was shown that this method can eliminate the interferences from external noises,and improve the accuracy of PETN detection.