In order to make balance between the effect and time consumption of image segmentation, an image segmentation algorithm based on feature fusion and cluster is proposed in this paper. Firstly, the segmentation granularity is obtained adaptively according to the coarseness of image; secondly, different features are extracted, multi features are fused and classified by K-means clustering, and the image can be segmented quickly. Experiments show that the proposed algorithm can segment image quickly, and the segmentation effect is good, which has verified the validity.