Objective The damaged regions of current image inpainting algorithms are mainly determined manually, and cannot identify the damaged regions automatically, leading to low repairing efficiency. This type of algorithms repairs the image by extracting the non-continuous edge of the pixel missing regions, which often causes visual discontinuity of the reconstructed image, and the inpainting path is always randomly determined, which increases the time consumption of the algorithm. In this paper, an automatic image inpainting algorithm based on topological gradient minimum inpainting paths coupled fuzzy C-mean was proposed. Methods The goal of automatically identifying the damage area was achieved by introducing the fuzzy C-mean to compute the membership function of the distance between pixels and cluster center in the damaged regions. Then the continuous edge of the pixel missing region was obtained by embedding topological gradient, defining the key point selection law of the pixel missing region and establishing weight distance function; the reconstruction path with lowest inpainting path cost was designed and the image was reconstructed. Finally, the image inpainting feedback mechanism was constructed based on PSNR to optimize the inpainting image. Results The simulation results showed that comparing with other image inpainting algorithms, the proposed algorithm could automatically identify the pixel missing region and extract the continuous contour of the pixel missing region. Meanwhile, this inpainting algorithm had a better inpainting effect and higher efficiency to eliminate the blurring effects and visual discontinuity. Conclusion The proposed algorithm could realize automatic image inpainting, and could improve the quality and efficiency of the repaired images.