The invention discloses a three-dimensional magnetic flux leakage detection defect reconstruction method based on a generative adversarial network, and belongs to the technical field of nondestructivetesting. The method is characterized by comprising the following steps: step 1, acquiring a defect image by using a depth camera to obtain defect data; 2, adopting a magnetic flux leakage detection instrument, and obtaining magnetic flux leakage data through post-processing of the magnetic flux leakage data; 3, dividing all data into sample data and test data, dividing the sample data into training data and verification data; 4, matching the magnetic flux leakage data in the sample data with defect data; 5, obtaining a GAN final model; and step 7, obtaining a defect three-dimensional contour.According to the three-dimensional magnetic flux leakage detection defect reconstruction method based on the generative adversarial network, three-dimensional reconstruction is carried out on the bottom plate by using the GAN model, defect three-dimensional contour reconstruction can be rapidly carried out, the reconstruction speed is high, the stability is good, the precision is high, the robustness is good, and the reconstruction speed and precision of various irregular defects are facilitated.