The invention discloses an SAR image sparsing denoising method based on change detection. The SAR image sparsing denoising method includes the steps of firstly, inputting an image; secondly, obtaining a changed area image; thirdly, setting a change classification; fourthly, obtaining a two-classification binary image; fifthly, obtaining a two-area image; sixthly, conducting denoising on unchanged areas; seventhly, estimating noise standard deviation; eighthly, obtaining a synthetic image; ninthly, conducting sparsing dictionary denoising; tenthly, outputting a result. The denoising preprocessing is conducted for the SAR image change detection, and the SAR image sparsing denoising method has the advantages that the consistency of a changed area is kept well, the noise of an unchanged area can be well restrained, meanwhile, the texture detail information, outline detail information and edge detail information of an SAR image, especially of the changed area which has great influence on change detection at later period are reserved, scratches produced after denoising is conducted on the image are eliminated, the pseudo information produced when noise is restrained is reduced, and the accuracy of the change detection is improved.