Mobile Ad hoc Networks (MANETs) are susceptible to various node misbehaviors due to their unique features, such as highly dynamic network topology, rigorous power constraints and error-prone transmission media. Significant research efforts have been made to address the problem of misbehavior detection. However, little research work has been done to distinguish truly malicious behaviors from the faulty behaviors. Both the malicious behaviors and the faulty behaviors are generally equally treated as misbehaviors without any further investigation by most of the traditional misbehavior detection mechanisms. In this paper, we propose and develop a policy-based malicious peer detection mechanism, in which context information, such as communication channel status, buffer status, and transmission power level, is collected and then used to determine whether the misbehavior is likely a result of malicious activity or not. Simulation results il-lustrate that the policy-based malicious peer detection mechanism is able to distinguish malicious peers from faulty peers with high confidence. Moreover, the mechanism converges to a consistent view of malicious nodes amongst all the nodes with a limited communication overhead.