The vigorous development of new energy vehicles has brought new opportunities and challenges to related industries. In order to gain greater competitive advantages, major automobile manufacturers have begun to focus on using motor current ripple signals directly to achieve various functions in a cheaper and more integrated way. However, the current ripple signal generated when the motor rotates is mixed with a lot of noise. If it is to be put into practical application, it needs to be denoised first, such as the improved wavelet threshold denoising method studied in this paper. First, the denoising principle and the shortcomings of the traditional method are introduced, and the design of the improved method is given on this basis. Subsequently, verification is conducted using simulated signals and actual current ripple signals, with SNR (Signal to Noise Ratio) and RMSE (Root Mean Square Error) used as quantitative evaluation criteria for comparing and analyzing the results. Finally, the experiments have confirmed that the improved threshold method proposed in this paper outperforms traditional methods in noise removal, proving its significance and value for further research.