The use of materialized views in commercial database systems and data warehousing systems is a common technique to improve the query performance. In past research, the view selection issue has essentially been investigated in the centralized context. In this paper, we address the view selection problem in a distributed scenario. We rst extend the AND-OR view graph to capture the distributed features. Then, we propose a solution using constraint programming for modeling and solving the view selection problem under multiple resource constraints in a distributed context. Finally, we experimentally show that our approach provides better performance resulting from evaluating the quality of the solutions in terms of cost saving.