Chinese folk paper-cutting is the crystallization of the wisdom of the working people, and a cultural treasure that has been passed down from ancient times to generations. This article mainly studies the design style changes of folk paper-cut patterns under the influence of visual art. Before proceeding to the pattern transfer mode, it is necessary to preprocess the most similar paper cuts in the input image and paper cut library. After drawing the paper-cut outline, extract the appropriate pattern from the pattern library and embed it in the paper-cut image according to the characteristics of the folk paper-cut. At the same time, the position and direction of the pattern need to be determined. Finally, pattern recognition is performed. Binarize the image by closed value segmentation, clearly distinguishing background and object. Use four adjacent edge detection algorithms to obtain the edges of the image. By tracking the edge of the image, adjust the coordinates in a counterclockwise or clockwise direction in pixels and save them. The data shows that the algorithm used in the experiment has a good discrimination effect for different types of patterns. Its total recognition rate is 91.06%. The results show that the multi-scale attributes of wavelet moments not only have the overall characteristics of the image, but also have local characteristics, so as to better characterize the difference in image details, and the recognition rate is higher when identifying similar objects.