An SAR image change detection method based on context salience detection and SAR image change detection is disclosed. Two phase SAR images of the same region are input; logarithmic ratio difference chart is calculated; significance Region of Difference Map Based on Context significance detection algorithm is extracted; according to the salient area mask of the difference map, the background of thetwo temporal SAR images is masked, and the SAR image after the masking is obtained; based on fuzzy C-means clustering, the change region of SAR image after the mask is extracted, and the training samples of pseudo-label are obtained; multi-layer sparse automatic encoder is trained by using masked SAR images and pseudo-label training samples; the final SAR image change region is extracted by the trained network. The invention can effectively overcome the influence of the speckle noise on the SAR image change detection, meanwhile, a multi-layer sparse automatic encoder is designed, the effective change characteristics of the SAR image are extracted, the accuracy of the change detection can be improved, and the invention can be used in the technical fields of multi-temporal SAR image changedetection and the like.