The primary obstacle typically encountered in Mobile Ad hoc Networks (MANETs) pertains for mitigating the impact of attacks prompted through malevolent nodes or identify promptly as well as addressing the certain nodes presence. This paper presents a hybrid technique in assault detection present in the context of MANETs. The research work focuses on addressing the challenge present in the MANET by introducing the robust Intrusion Detection System (IDS) using hybrid Machine Learning (ML) methods. The proposed approach for identifying attacks involves the utilization of the Adaptive Neuro Fuzzy Inference System in conjunction with the Rat Swarm Optimization Algorithm (ANFIS-RSOA). Hence, this hybrid ML approach has capability to produce high secure with precise and reliable outcomes. The suggested protocols concentrate on the security within a network by effectively identifying and mitigating potential assaults. The suggested methodology is executed within the NS2 platform and afterwards compared to various conventional methodologies, namely (PSO) Particle Swarm Optimization, (WOA), Whale Optimization Algorithm and Grey Wolf Optimization Algorithm (GWO). In order to evaluate the efficacy of the suggested methodology, it is subjected to testing using two distinct types of attacks, namely the (BHA) Black Hole Attack and the Wormhole Attack (WHA). This proposed ANFIS-RSOA method performance metrics such as jitter, throughput, delay, Packet Delivery Ratio (PDR), and low end-to-end delay is evaluated and compared with existing IDS methods. Moreover, the purpose of study is to protect both individual network nodes and their connections to one another.