The reliability of peers is very important for safe communication in peer-to-peer (P2P) systems. The reliability of a peer can be evaluated based on the reputation and interactions with other peers to provide different services. However, for deciding the peer reliability there are needed many parameters, which make the problem NP-hard. In this paper, we present two fuzzy-based systems (called FBRS1 and FBRS2) to improve the reliability of JXTA-overlay P2P platform. In FBRS1, we considered three input parameters: number of interactions (NI), security (S), packet loss (PL) to decide the peer reliability (PR). In FBRS2, we considered four input parameters: NI, S, PL and local score to decide the PR. We compare the proposed systems by computer simulations. Comparing the complexity of FBRS1 and FBRS2, the FBRS2 is more complex than FBRS1. However, it also considers the local score, which makes it more reliable than FBRS1.