Capillary electrophoresis (CE) is a powerful analytical tool in chemistry. Thus, it is valuable to solve the denoising of CE signals. A new denoising method called MWDA which employs Mexican Hat wavelet is presented. It is an efficient chemometrics technique and has been applied successfully in processing CE signals. Useful information can be extracted even from signals of S/N =1. After denoising, the peak positions are unchanged and the relative errors of peak height are less than 3%. It is an efficient chemometrics technique and has been applied. It is an efficient chemometrics technique and has been applied Useful information can be extracted even from signals of S / N = 1. After denoising, the peak positions are unchanged and the relative errors of peak height are less than 3%.