The purpose of image optimization is that the salient region of image is stand out, and is layout an image met the demand. The existed image optimization method is summarized, and their advantages and disadvantages respectively analyzed. The improved visual attention model is described, and is used to detect the salient region of image. The parameters of image optimization are calculated by the optimization function of image layout rules. At last, the weights of the images optimized parameters are adjusted by the image layout and adjustment function based on image optimized parameters, to achieve image layout and obtain the most optimal image, to realize image optimization. Experimental results validate that this methods not only achieve image optimization, but also accurately and automatically achieve to shorten distance and enlarge salient regions, improve the quality of the image optimization, and has good robustness.