The emergence of the Internet of Things (IoT) has transformed our daily experiences, providing improved connectivity, accessibility, and convenience. This interconnectedness, while beneficial, also presents challenges, as it creates a network vulnerable to various attacks. Designing an effective Intrusion Detection System (IDS) for IoT networks is intricate, mainly due to the sheer volume of data and the diverse IoT device landscape. Conventional IDS methods often falter in real-time data handling and analysis. Consequently, there’s an escalating need for sophisticated IDS solutions employing Machine Learning (ML) or Deep Learning (DL) techniques. This research delves into IoT network intrusion detection, harnessing the insights from the widely recognized BoT-IoT dataset. The primary objective is to bolster intrusion detection accuracy by amalgamating genetic algorithm (GA)-driven feature extraction and ensemble machine learning (EM) strategies. Feature extraction plays a pivotal role in IDS, striving to trim down data dimensions while preserving pertinent details. Genetic algorithms, revered for their prowess in optimization, are harnessed to pinpoint an optimal feature subset amplifying IDS discrimination capabilities. A novel Framework is introduced, synergizing GAs with a medley of ensemble ML algorithms, encompassing random forests, AdaBoost, Extra-Trees, XGBoost, and stacking methodologies. The Genetic Algorithm (GA) identifies a specific subset of features from the BoT-IoT dataset. Subsequently, ensemble Machine Learning (ML) models are trained and assessed, and their accuracy is determined based on these chosen features.