In this paper, a new diffusion coefficient is proposed, and a new partial differential equation denoising model is obtained. The new model can not only remove the Gaussian noise, but also remove the saltpepper noise. Meanwhile, the discretization scheme of the Laplace operator is improved, and more image information can be contained. Thus, the image features can be judged more accurately. The images processed by the new model do not incur the blocky effect which is widely seen in the images processed by the secondorder nonlinear diffusion. Compared with other fourthorder partial differential equations, the images do not generate speckles. Therefore the visual effect of the images is superior to that processed by the  YouKaveh model. Finally, the validity of the new model is proved by experiments.