Searching for entities, especially people are one of the most common activities
of Internet users. The problem of name ambiguity causes the search results for a
personal name to be a mix of web pages about different people sharing the same name. In this paper, we address the problem of web people search clustering, and propose a new hybrid approach to cluster web pages based on their association to different people. The key difference between the approach we propose and the existing techniques is that our approach analysis not only people attributes but also social relationships to improve the clustering quality. Our approach integrates social relationships with people attributes and maps them into an undirected attribute relationship graph. Then a graph clustering algorithm is used to group web pages into a set of disjoint clusters. The experiments over two real datasets indicate that our approach carries out better results than the counterparts.