In recent years, as one of the biometric identification technology, palm-print identification has received many reseachers’ attention. To solve the key problem of palm-print recognition – feature extraction, we propose a new method, which based on wavelet transform and principal component analysis. In general, we use wavelet transform to deal with palm print images and extract high-dimensional wavelet energy features, then reduce the dimensionality of high-dimensional wavelet energy features through principal component analysis, and remain the original feature energy maximally. The features extracted by this method not only reflect palm-print images’ information maximally, but also achieve the goal of data dimensionality reduction. Experiments show, the correct recognition rates of new method are much higher than those traditional methods such as LDA [1], PCA [2], 2DPCA [3], ICA [4] and so on.