The invention discloses a robust automatic image salient region segmenting method. The robust automatic image salient region segmenting method includes the steps of firstly, subjecting an image to superpixel segmenting to obtain geodesic distances, generating regions, boundary length and boundary communication values of image superpixels; secondly, subjecting an original input image to superpixel segment by an HSLIC (hexagonal simple linear iterative clustering) method, performing overall saliency detection by an SC (superpixel contrast) method to obtain a salient image of the image; finally, adopting the boundary communication values and a saliency value of the salient image, performing image segmenting and outputting a salient region segmenting result of the image. The robust automatic image salient region segmenting method has the advantages that the boundary communication values, the image saliency value and the image segmenting method are combined, and the image salient region segmenting result can be obtained by the aid of an automatic image segmenting technology.