Abstract: In order to overcome the limitation of single image feature segmentation, an image segmentation method based on the fusion of multiple features and spectral clustering was proposed. Firstly, the image was segmented by multi-feature based on the granular computing, and then the result was mapped to the hypergraph. Finally, the final segmentation result was obtained by using the spectral clustering algorithm. The experiment results indicated that the proposed method can obtain a relative better segmentation results compared with the Gpb (Globalized probability of boundary) algorithm.