The rapid expansion of the Internet of Things (IoT) and Cyber-Physical Systems (CPS) has revolutionized modern industries, smart environments, and critical infrastructure. However, this growth has also exposed these systems to a wide array of sophisticated cyber threats. Traditional Intrusion Detection Systems (IDS) often fall short in addressing the dynamic and resource-constrained nature of IoT and CPS environments. This paper presents an intelligent IDS framework that integrates advanced Machine Learning (ML) and Deep Learning (DL) techniques to detect malicious activities with improved accuracy and responsiveness. The proposed system incorporates key components such as data preprocessing, feature engineering, and a dual-layered detection engine, tailored for heterogeneous and real-time scenarios. Extensive experiments are conducted using benchmark datasets including NSL-KDD, CICIDS2017, and IoT-23 to evaluate the model’s performance. The results demonstrate superior detection rates, reduced false alarms, and effective handling of small-sample and imbalanced attack data. Additionally, real-world deployment considerations and use-case evaluations in smart homes, healthcare, and industrial CPS environments are discussed. This study emphasizes the potential of intelligent IDS as a robust defense mechanism for securing next-generation connected systems.