P2P reputation systems are useful to evaluate the trustworthiness of peers and to combat malicious and selfish peer behaviors. The reputation system assigns each peer a global reputation scores through collecting locally generated peer feedbacks and aggregation. In traditional reputation system, a peer will rate the other by a score, after finished a transaction, No more detail of transaction are considered in the process. In order to solve this problem, we propose a novel reputation aggregation scheme, detail of transaction could be considered, based on set pair analysis (SPA). The empirical evaluation results reveal that the proposed approach is scalable, accurate, robust, and fault-tolerant.