The purpose of this paper is to extract data features and denoise the interference excitation source in vehicle electromagnetic compatibility test. The lifting wavelet packet algorithm inherits the multi-resolution characteristics of the classic (first generation) wavelet transform. The transform is only carried out in the time domain, which can achieve in situ operation. It has the advantages of small space occupation, fast transformation speed, easy inversion, etc. It can use energy conservation criteria to extract characteristic energy to identify the conducted interference sources in the vehicle, and the obtained characteristic spectrum is used as the modulation array of the excitation source of the vehicle numerical simulation. In this paper, the collected interference signals are decomposed into lifting wavelet packets, and then the characteristic energy is extracted to identify the conducted interference sources in the vehicle. Signal to noise ratio (SNR), root mean square error (RMSE) and peak error (PE) are used to verify the consistency between the simulation signal and the original signal. The results show that the lifting wavelet packet algorithm has a strong ability to identify the conducted interference sources in the vehicle.