Image search reranking is one of the effective approach to refine the text-based image search result. Textbased image retrieval suffers from essential problems that are lead to the incapability of the associated text to appropriately evoke the image content. In this paper, reranking methods are put forward to address this drawback in scalable fashion. Based on the classifiers for each and every predefined attributes,each and every image is represented by an attribute feature consisting of the responses from these classifiers. This hypergraph can be used to model the relationship between images by integration of low-level visual features and attribute features. Hypergraph ranking is then performed to order the images. Its basic principle is that visually close images should have identical ranking scores. It improves the performance over the text-based image search engine.