In Peer-to-peer (P2P) systems, there has been little work on managing data replication in the presence of updates. However, important P2P applications that involve collaboration require multi-master replication support. In this paper, we adopt optimistic replication and propose a novel distributed algorithm to reconcile conflicting updates in the context of APPA1 (Atlas Peer-to-Peer Architecture) data management system. APPA has a network-independent architecture that can be implemented over various P2P networks. Our algorithm for distributed semantic reconciliation (DSR) enables multi-master replication and assures eventual consistency among replicas. We validated the DSR algorithm through implementation and simulation using Java and SimJava. The experimental results show that DSR has good performance and scale up.