Regularization is popular in image restoration recent years. We analyze the preconditioning conjugate gradient method with Tikhonov regularization under the periodic boundary conditions, and propose a new preconditioning matrix and the varying regularization parameter method. At first, we choose a larger regularization parameter to restrain the noise in the restored image, get a convergent result to modify the initial gradient. After that, we choose a smaller one to increase the details. Experiments on a set of image restoration and reconstruction benchmark problems show that the proposed algorithm performs favorably in comparison with several state-of-the-art regularization image restoration algorithms.