The advent of the Internet of Things (IoT) has significantly altered human interaction with the environment, enabling smart ecosystems that simplify day-to-day activities. Despite these benefits, the expansion of IoT technology has also escalated security and privacy vulnerabilities, necessitating robust network protection mechanisms. This paper introduces an innovative approach that combines Mutual Information Feature Selection (MIFS) with the Flower Pollination Algorithm (FPA) and Particle Swarm Optimization (PSO) for effective feature selection. Furthermore, the detection task is performed using the Light Gradient-Boosting Machine (LightGBM). However, the empirical tests on the IoTID20 dataset reveal that the proposed methodology surpasses various stateof-the-art intrusion detection techniques in accuracy, recall, and F1-score. Moreover, the approach exhibits lower false positive rates and higher detection rates, affirming its efficacy in identifying IoT network intrusions.​