This paper addresses the problem of supporting ranked join in peer-to-peer (P2P) networks. Ranked queries produce results that are ordered by a certain computed score. Thanks to the ability of answering information retrieval style queries, ranked query has become a critical need for many applications relying on relational databases, where users do not expect exact answers to their queries, but instead a rank of the objects that best match their preferences. In this paper we propose a novel algorithm P Join for supporting efficient ranked join queries in P2P networks, which is a part of our work to introduce database query processing facilities in P2P networks. The existing ranked join query algorithms consume an excessive amount of bandwidth when they are applied directly into the scenario of P2P networks. By pruning irrelevant tuples before join probing, PJoin reduces considerable amount of bandwidth cost. Performance of the proposed algorithm is validated by extensive experiments.