Wavelet transform is widely applied in signal processing,which can analyze time domain or frequency domain information.But the traditional wavelet transform depends on Fourier transform,and its realization is based on a large amount of convolution computation,leading to the low operation speed.This paper discusses the principles of second generation wavelet transform,and applies it to process the electroencephalo-graph (EEG).Lifting scheme is the key technique to construct the second generation wavelet transform.High-frequency information is determined by predicting,and the correct low-frequency information is obtained through update.Lifting scheme does not depend on the Fourier transform,thus greatly improves the speed of operation.By analyzing the basic principles of lifting scheme,we extracted the four rhythms of EEG through second generation wavelet transform,and got satisfactory results.