Abstract: This paper proposes an efficient codebook design algorithm for image compression. We firstly sort the training vectors using principal component analysis to reduce the computational complexity,and then make full use of the near global optimal searching ability of genetic algorithm to compute the codebook. The experimental results show that our algorithm outperforms the popular LBG algorithm in terms of computational requirement,and even when the codebook size is as small as less than 64,our algorithm still shows its advantage.