The invention discloses a salient region detection method based on local self-adaptation regression and belongs to the technical field of image processing. The method includes the steps that firstly, a Lab image is extracted; secondly, feature descriptors are generated; thirdly, effective features are extracted, and single-scale salient regions are generated; finally, salient region merging is conducted. According to the salient region detection method based on the local self-adaptation regression, a good effect can be obtained when the natural image salient areas with obvious noise are detected, calculating is simple and convenient, and the method is high in efficiency and good in implementability.