Trust is multi-faceted and the peer’s needs are different in different situations. A peer may need to consider its trust in a specific domain of another peer’s capability or in a combination of multiple domains. Current Peer-to-Peer trust model could not promise the trust computation of different domains. This paper presents a novel Peer-to-Peer multiple-grain trust model and gives a distributed implementation method, which considers the trust computation of different domains. Mathematical analyses and simulations show that, compared to the current trust model, the proposed model is more precise on trust computation of multiple domains and more robust on trust security problems.