The high count of IoT device and poorer security are extremely probable to be snatched that resulted in DDOS attacks. By depleting the node battery or disrupting the signal, this assault taints the network’s overall accessibility or the accessibility of specific nodes. This article introduces a DL-oriented approach to recognize such attacks and mitigate them. This article proposes a new, four-stage assault detection system for IOT. Data normalization is initially carried out during pre-processing. Then, “enhanced Recursive Feature Elimination (RFE), and improved second order technical indicator based features (ATR, CMF, CTI, and improved EMA) are extracted,” along with “higher order statistical features (kurtosis, variance, skewness). The best features were also chosen using the CMIHBO method. The existence of attacks is then determined by averaging the outputs of a group of classifiers, including “Recurrent Neural Network (RNN), Bidirectional Gated Recurrent Unit (BI-GRU), and Bidirectional Long Short-Term Memory (BI-LSTM).” With Cat and Mouse Integrated HBO, the Bi-LSTM weights in particular are calibrated to perfection (CMIHBO). “Recommended BAIT focused mitigation” is applied to eliminate the offender nodes from the networks if an incident is discovered. In the end, many measures are established to improve the deployed approach.