The project focuses on bolstering threat detection efficiency within Internet of Things (IoT) systems through an intelligent approach. IoT systems, comprising devices, sensors, networks, and software, often struggle with security vulnerabilities exploitable by attackers. Using machine learning algorithms and principal component analysis (PCA), the study targets the identification of DDoS attacks, a prevalent menace to IoT systems. Principal component analysis aids in data dimensionality reduction, streamlining datasets while preserving critical information. Evaluation encompasses metrics like accuracy, precision, recall, and F1-Score to gauge model performance accurately. Employing CICIDS 2017 and CSE-CIC- IDS 2018 datasets, the models are rigorously trained and tested. The proposed approach exhibits superior performance and diminished training time compared to prior methodologies, showcasing its efficiency in bolstering threat detection within IoT systems. The study further enhances that the project integrates ensemble techniques such as Voting Classifier (RF + Ada boost) and Stacking Classifier (RF + MLP with LightGBM), culminating in a refined and precise predictive model achieving 100% accuracy. This research not only advances threat detection capabilities but also highlights the potential of ensemble methods in fortifying IoT system security.