Abstract: To improve the recognition rate, a palmprint recognition method based on multi-dimensional ensemble empirical mode decomposition (MEEMD) and two-dimensional principal component analysis (2DPCA) is proposed in this paper.Palmprint images are decomposed with MEEMD to get IMF components, then reconstruct the palmprint images with the high frequency IMF components to get the recognition palmprint images set.As the last step, the reconstructed palmprint set is input to 2DPCA to recognize.The reconstructed palmprint images have more high frequency characteristic details than the original palmprint images, so its recognition rate is higher.The palmprint database of Hong Kong Polytechnic University is employed in experiments.The results show the higher recognition rate and faster recognition speed of our method.