This paper proposes a network intrusion detection system (NIDSE) for smart home devices using an ensemble model, specifically XGBoosting, to detect network attacks. Smart home systems are increasingly popular for their convenience and automation, but they also pose significant security risks, especially with vulnerabilities in devices like baby monitors and IoT cameras. The Mirai malware attack, which targeted IoT devices, highlights the need for robust intrusion detection mechanisms. The NIDSE system is designed to monitor smart home devices passively and detect intrusions by classifying attacks using XGBoosting, an ensemble method that improves model performance by sequentially adding models to correct errors. The system was tested on the IoT network intrusion (IoT-NI) dataset, which includes various types of network attacks such as host discovery, SYN, ACK, and HTTP flooding. The results show that the XGBoosting classifier achieved a micro average precision of 94% and macro average precision of 85% in classifying nine types of attacks. The system outperformed other classifiers like Random Forest and Balanced Random Forest in terms of precision and accuracy. The NIDSE system provides endpoint security for smart home devices by using XGBoosting to enhance attack detection, improving accuracy, precision, and recall. The system is implemented on edge devices, reducing latency and bandwidth issues. However, it cannot detect clone attacks in the network. The study also reviews existing intrusion detection systems and compares the performance of different machine learning algorithms for network intrusion detection. The proposed NIDSE system is an effective solution for detecting network attacks in smart home environments.This paper proposes a network intrusion detection system (NIDSE) for smart home devices using an ensemble model, specifically XGBoosting, to detect network attacks. Smart home systems are increasingly popular for their convenience and automation, but they also pose significant security risks, especially with vulnerabilities in devices like baby monitors and IoT cameras. The Mirai malware attack, which targeted IoT devices, highlights the need for robust intrusion detection mechanisms. The NIDSE system is designed to monitor smart home devices passively and detect intrusions by classifying attacks using XGBoosting, an ensemble method that improves model performance by sequentially adding models to correct errors. The system was tested on the IoT network intrusion (IoT-NI) dataset, which includes various types of network attacks such as host discovery, SYN, ACK, and HTTP flooding. The results show that the XGBoosting classifier achieved a micro average precision of 94% and macro average precision of 85% in classifying nine types of attacks. The system outperformed other classifiers like Random Forest and Balanced Random Forest in terms of precision and accuracy. The NIDSE system provides endpoint security for smart home devices by using XGBoosting to enhance attack detection, improving accuracy, precision, and recall. The system is implemented on edge devices, reducing latency and bandwidth issues. However, it cannot detect clone attacks in the network. The study also reviews existing intrusion detection systems and compares the performance of different machine learning algorithms for network intrusion detection. The proposed NIDSE system is an effective solution for detecting network attacks in smart home environments.