In order to improve the compression rate of high-quality image, the Image Compression Algorithm based on psychological visual redundancy and Principal Component Analysis was proposed. According to the principle of various types information with different significance degrees in the human visual process, the unimportant psychological visual redundant information was eliminated under it did not affect the perceived quality of the image, reducing the dimension of image that eliminated mental visual redundancy with PCA which made a progress in image compression. The experimental results show that the proposed new compression algorithm makes the image compression ratio increased by 1 times under it does not affect the perceived quality of the image and greatly reduces the storage space.