Abstract: Different orders of vanishing moments of the wavelet, the amplitude-frequency curve of the corresponding filters may be different too, so using different wavelet filters for de-noising on the ECG signal, the denoising effect will be significantly different. In this paper, we study the amplitude frequency characteristics of wavelet filter by building the mathematical model and clarify the relationship between the wavelet amplitude characteristic and the order of the vanishing moment of the wavelet filter, and try to provide the theoretical basis for selecting the optimal wavelet filter. In this paper, we proposed a special method to realize the accurate notch denoising based on the frequency characteristics of ECG noise, which effectively preserves the singularity and feature of the signal and reduces the signal distortion. The experimental results show that the lifting wavelet filters with the relative optimal order of the elimination moment can make the signal energy distribution more concentrated and the denoising effect more better.