In order to further improve the de-noising ability of the conventional polynomial matching algorithm, an improved polynomial matching filtering algorithm based on edge protection is proposed for the deficiency of the conventional polynomial matching algorithm and the image characteristics and noise problems. Based on the conventional polynomial matching algorithm, this method improves the selection method of the filter window, extracts the adaptive sliding filter window along the direction of the image texture, and selects the window with the smallest matching error for matched filtering and uses it as the final output result. Then add Gaussian white noise and pepper and salt noise to the gray image and CT image respectively for testing. The data verification shows that the method can maintain edge / texture information under the premise of effectively suppressing noise, and the peak signal-to-noise ratio is increased by more than 80% on average, mean square The root error is reduced by more than 80%. It is concluded that compared with the conventional polynomial filtering method, median filtering method, bilateral filtering method and edge preserving filtering method, the improved method can effectively improve the image visual effect, meet the image application requirements, and has good application prospects.