The social search use social ability to participate in the screening and sorting of search results in order to make the result to tend to reflect the hopes of the people better. Based on the prototype of social search, the concept of user tag feedback scores is employed. Based on this concept, a tag-based feedback web ranking algorithm is designed. The algorithm can efficiently use the user's feedback. Finally, a variety of evaluation methods are used to analyze the search results of different sort. The experiments show that the new sorting algorithm can meet the user's real needs better.