Abstract: Query rewriting technique based on views is an important solution to sharing data in PDMS (peer data management system). An extended Datalog technique based on peer constraints is proposed to optimize query processing in PDMS. In particular, the query optimization problem in data-intension and logical tight-coupled PDMS environment is studied. The definition language for PDMS views, PView and the definitions of local and global peer constraints are proposed, as well as the query rewriting optimization algorithms. A cost model is given to evaluate-the queries in PDMS. Experimental results show that the proposed approach outperforms other view-based query rewriting approaches.