This paper presents a wavelet based Kernel Principal Component Analysis (KPCA) palmprint recognition method for human identification. The intensity values of palmprint images are first normalized by using their mean and their standard deviation. The normalized images are then transformed to the spectral domain by using wavelet transform and lowest frequencies are selected by filtering. Next, the feature vectors are formed with KPCA method which divergences samples on the nonlinear space. Finally, weighted Euclidean distance based nearest neighbor method is realized for palmprint classification. Experiments are performed on the most-well known public palmprint database, PolyU, includes 600 samples of 100 different persons.