The invention provides an image compression method. The image compression method includes the following steps that a data matrix of an image to be compressed are generated, and the data matrix is centralized or standardized; a variance matrix of the centralized or standardized data matrix is calculated; a characteristic polynomial of the variance matrix is converted into a high-order characteristic polynomial, and the number of roots of the high-order characteristic polynomial is judged; according to the number of the roots and a preset initial solution, iteration solving is conducted on the high-order characteristic polynomial, when the number of the roots acquired through iteration solving is four, the four roots are calculated according to a mathematical expression, acquired through the current iteration solving, of the characteristic polynomial, all characteristic roots are output, and characteristic vectors are calculated according to the characteristic roots; a transformation matrix is acquired according to the characteristic vectors, and the compressed image is acquired by multiplying the transformation matrix by the data matrix. By the adoption of the image compression method, the operand in the image compression process is small, and the compression speed is high.