<p>MANET security is a difficult process that so many researchers in security working on networks continuously seek to resolve. However, different Intrusion Detection System (IDS) frameworks have been developed by traditional research to provide security and reliability in MANET. The computational operations remain complex, and training time becomes long; thus, we have reduced convergence speed and output errors as major problems of the system. In this paper, a smart and intelligent intrusion prevention framework for intrusion detection within the network is introduced. In the intrusive schemes, these two operational phases function as intrusion detection and intrusion prevention modules to increase the capability to process larger quantities of IDS raw data in a shorter time frame. Intelligent optimization is used in the process of detection of the intrusions combined with classification models that identify individual intrusions from the network. After authenticating trusted users in the network, the work of the Trust Score Evaluation (TSE) model begins to make a communication. Using provided datasets, distance value is used to establish attribute clusters in preprocessing as done by the Markovian Data Clustering Model (MDCM). It represents the prime purpose because it gives a better control of dataset value ranges. After clustering the data into clusters by the Markovian Data Clustering Model (MDCM), the BE2GO algorithm optimizes the features to train the classifier efficiently and more accurately with simpler detection methods for the classification. The optimized feature set is employed to classify attack categories through the Stochastic Elman Neural Network (SENN) algorithm with intelligent features. Comparative results are evaluated of the BE2GO-SENN simulation with several assessment indicators.</p>