The invention discloses a method for repairing and reproducing old films based on deep learning, which includes the following steps: S1, extracting the frames of a video through ffmpeg, and forming training data sets of a de-interlacing scanning model, a video frame interpolation model, a de-blurring network and a super-resolution model respectively; S2, training a de-interlacing scanning networkmodel; S3, training a video frame interpolation network model; S4, training a de-blurring network; S5, training a super-resolution network; and S6, training a de-noising network. In the invention, oldfilms are repaired by using de-interlacing scanning, video de-noising, video de-blurring, video frame interpolation and super-resolution technologies based on deep learning. Compared with manual operation, the stability is higher, the computing speed is improved, and the accuracy of image restoration is improved. The method has the advantages of good restoring effect of processed images, high clarity of restored images, convenience of use, low cost, and the like.