对等网信誉系统的一个关键问题就是信誉评估的准确性,一般通过考察推荐方的反馈可信度来给其反馈加权.提出一种新的反馈可信度评估模型,模型基于推荐方历史推荐行为满意度预测其未来的推荐行为,模型提出的“虚检验”机制加速了评估的过程,并且提高了准确性.仿真和分析表明,模型能有效抑制恶意推荐行为,降低其对最终信誉评估的影响. A key problem of P2P reputation system is the accuracy of reputation assessment, which is usually weighted by examining the feedback credibility of the recommendation parties.A new evaluation model of feedback credibility is proposed, which is based on the historical recommendation of the recommendation party Behavior satisfaction predicts its future recommendation behavior, the “virtual test” mechanism proposed by the model accelerates the evaluation process and improves the accuracy.The simulation and analysis show that the model can effectively suppress the malicious recommendation behavior and reduce its impact on the final reputation Impact of the assessment.