The rapid growth of Internet of Things (IoT) devices presents various challenges, notably in security, speed, accessibility, and scalability. With IoT’s expanding reach, devices become more interconnected, leading to increased vulnerability to malicious exploitation. Consequently, more individuals face susceptibility to attacks. Addressing these escalating security risks, this paper adopts a proactive stance, identifying potential vulnerabilities in IoT applications exploited by malicious actors. Protecting IoT devices proves daunting due to their complex functionalities. To combat this, the paper introduces a novel approach integrating supervised learning, K-Nearest Neighbors (KNN), and the Constant-Time Ensemble Learning Classifier (CTELC) algorithm. This method aims to enhance the security of software-defined IoT (SD-IoT) networks by mitigating Distributed Denial of Service (DDoS) attacks promptly and accurately. Comparing its efficacy with other supervised machine learning algorithms, the proposed approach, utilizing CTELC with NCA feature selection, achieves a significant 66% reduction in detection time while attaining an impressive 97.7% accuracy in identifying DDoS attacks.%, achieving an impressive accuracy of 97.7% in identifying DDoS attacks.