The laser acoustic signal has the characteristics of pulse width and frequency width. In a certain range, with the increase of laser energy, the energy distribution in frequency domain remains stable. Wavelet packet technology is used to analyze the acoustic signal characteristics of the target echo signal. In this paper, db4 wavelet base is used to decompose laser acoustic signals. The energy characteristics of decomposed signals are extracted and the energy distribution characteristics of acoustic signals are analyzed. In order to determine the time domain characteristics of the target laser sound signal before and after reflection, the reconstructed signals of different nodes after decomposition are reconstructed, and the correlation analysis between the reconstructed signal and the original signal is conducted to determine the effective filtering frequency band of the signal. The data analysis shows that the wavelet packet analysis method can effectively analyze the transient characteristics of laser acoustic signal, and select the signal filtering frequency band according to the energy characteristic value to effectively filter the signal, which can provide a reference for the research of underwater acoustic target identification.