Abstract: In the image denoizing process, the anisotropic diffusion algorithm is prone to image edge information loss and texture detail blur. Considering this, an image denoizing algorithm combining improved PM and recursive filter is proposed. This algorithm first preserves the edge information of the image by controlling the coefficients of the diffusion in the gradient direction and that perpendicular to the gradient direction, and then the recursive filtering algorithm is used to reduce the number of iterations to protect the detailed texture features of the image. The experimental results show that the new method has superior effects in image denoizing and edge detail information protection, and greatly reduces the mean square error. The peak signal-to-noise ratio of the image is also significantly improved.