When the ECG signal is digitally filtered, it is easy to cause ECG distortion due to its amplitudefrequency characteristic and transition zone. In order to make the original noisy ECG signal have the same smoothness and minimum variance as the ECG signal after noise reduction It is analyzed and analyzed by using the data of MITBIH database and the measured ECG data respectively. The simulation results show that the wavelet denoising algorithm is used to denoise the wavelet denoising algorithm. The algorithm can be used to denoise the ECG signal, get the original ECG signal with clear distortion, Signal to noise ratio SNR of 30 or more, mean square error MSE below 0.1, the design method is simple, the calculation is small, Good application prospects and extensive use of value.