In order to improve the visual effect and detail expression of the image, a method of image multi-frequency detail feature enhancement based on multi-scale filtering is proposed. The gray level transformation and denoising are applied to the image, and the bilateral filter is used to divide the preprocessed image into a low-frequency background layer image and a high-frequency detail layer image, covering the multi-frequency detail features from contour to texture. For the image enhancement processing of two scales, the two are fused by the fusion method, and the complete detail enhanced image is obtained. The experimental results show that the information entropy and the average gradient data of this method are higher, the information entropy is 5.969 5 bit, and the average gradient is 0.022 5. It exhibits good performance in multi-frequency detail feature extraction and enhancement, and has advantages in the richness of detail information and the clarity of detail edge.