The IoT is increasingly vulnerable to cyber attacks it widespread use of inexpensive, low-power devices that are distributed across various networks. These devices have limited computational resources, making them susceptible to malicious activity. As such, enhancing the security of the IoT ecosystem is of critical importance. One promising approach is to design lightweight Intrusion Detection Systems (IDSs) that can monitor and protect IoT devices against attacks while considering their limited resources. The objective of this study is to enhance the security of IoT device by introducing lightweight Intrusion Detection System based on machine learning techniques, particulerly feature selection and feature classification. IDS plays an important role in safeguarding Information and Communications Technology (ICT) systems by identifying and mitigating unauthorized or malicious activities. However, traditional IDS methods, which are typically designed for more powerful computing systems, are not always suitable for IoT applications. IoT devices are face limitations such as constrained memory, limited battery capacity, and specific protocol requirements, making it challenging to deploy conventional IDS solutions.