Internet-of-Things (IoT) connects various physical objects through the Internet and it has a wide application such as in transportation, military, healthcare, agriculture, and many more. Detecting attacks in IoT networks involves identifying abnormal patterns in device behavior or network traffic that indicate potential threats. This research enhances system security by enabling early detection. However, these networks face escalating cybersecurity threats that can cause security issues. To overcome these issues a novel FLAMingo based intrusion detection using deep learning netWORK (FLAM-WORK) has been proposed to identifying network traffic behavior and mitigating cyberattacks in IoT to provide guaranteed network security. The network traffic data packets are gathered from the Input Devices. The data are pre-processed to enhance the data quality. The proposed method utilizing the Convolutional Neural Network (CNN) for feature extraction to understand the complex data easier. Next Feature Selection done by Flamingo Search Algorithm (FSA) to enhance classification accuracy. A Ghost Net classification technique classifies the extracted feature into three classes such as Low, Normal, High. Measures including Specificity, F1-Score (F1S), Accuracy, Precision (PR) and Recall (RC) are used to assess the suggested approach. Compared to current models, experimental results using TON-IoT datasets show higher accuracy. The accuracy of the approach in the TON-IoT dataset is 1.3%, 0.7%, and 1.1% higher than that of the current BMEGTO-KNN, EBWO-HDLID and LIME techniques respectively.