Autonomous wireless infrastructure-less networks, i.e., mobile ad hoc networks (MANETs), enable effective communication in environments lacking infrastructure, such as emergency services and military operations. Despite that, network characteristics such as mobility and decentralization make it vulnerable to security attacks at the network layer. Malicious nodes intentionally drop the packets from the communication to significantly damage the network's performance. In a network, packets are also dropped by nodes unintentionally during the communication due to system fault or constrained resources. Existing packet drop mitigation mechanisms fail to accurately differentiate between malicious packet drop and unintentional packet drop, which leads to ineffective node classification and reduced network performance. The paper aims to classify packet-dropping behavior accurately with the help of the K-nearest neighbors (KNN) algorithm named as (KNN)-driven intrusion detection system (KNN-IDMS). KNN's ability to effectively recognize patterns helps to accurately classify whether the packet drop is due to malicious intent or unintentional. The proposed KIDS performance is computed using an NS2 simulator. The results indicate excellent packet delivery and energy efficiency improvements compared to existing methods. The proposed KNN-IDMS mitigates packet-dropping nodes and improves network performance, ensuring that it is an effective solution to enhance security.