Tag-based image search is an essential method to discover images shared by users in social networks. Social media Websites permit users to comment or explain the images with free tags, which can enhance the development of the web image retrievals. But measuring the top ranked image search is challenging. In this paper, the authors propose diverse ranking approach for tag-based image retrieval. First, tag graph based on the similarity between each tag is constructed; next the community detection method is carried out to extract community for each tag. Then next, inter-community and intra-community ranking are utilized to acquire the final retrieved results. In the inter-community ranking process, an adaptive random walk model is employed to rank the community based on the multi-information of each topic community. Beside this an inverted index structure is developed for images to accelerate the searching process. Experimental results on Flickr data set and NUS-Wide data sets show the effectiveness of the proposed and implemented methodology.