In order to inherit and protect the intangible cultural heritage, Fengxiang woodblock prints, in Shaanxi region and solve the problems of difficult color restoration of traditional handicrafts and low efficiency of pattern combination innovation, the work aims to combine the Generative Adversarial Networks and Fengxiang woodblock prints to promote its development in a multi-dimensional way. By means of high-definition scanning, adaptive threshold detection method, standard color value extraction and data augmentation, systematic generalization and standardized collection and storage were carried out to establish the dataset of Fengxiang woodblock prints, and CycleGAN algorithm was used to train the network model to complete the migration experiments. Through the combination of art and technology, both the color restoration of ink line drafts in the prints and the rapid coloring of innovative design patterns were completed. In the historical data restoration and application of Fengxiang woodblock prints, through the combination of modern design methods and color style migration function, the method can adapt to the modern development trend and complete the task of living inheritance while providing new ideas for similar research.