The wavelet transform has remarkable advantages and wide applications in denoising because of its characteristic of good time-frequency. Based on the analysis of traditional wavelet denoising methods, which are based on Fourier transform, an improved method is proposed. It overcomes the shortcomings of the traditional Fourier denoising method. In this paper, the denoising procedures are introduced respectively based on the wavelet transform and the method of connecting the wavelet threshold with the wavelet basis is adopted. Through Matlab simulation and concrete data, it arrives at the conclusion that the method of signal denoising based on the wavelet transform is obviously more effective and better than the traditional method based on Fourier transform.