The use of Internet of Things (IoT) technology has undergone a boom in prominence equivalent to the spread of smartphones. IoT devices are at high risk for security threats, including DDoS attacks from IoT botnets due to their small size and limited resources. Implementing strong security measures is essential to protect these devices and their sensitive data from cyber-attacks. Machine-learning algorithms used for the early identification of these dangerous botnets. In this research, we seek to identify Botnets for early detection. The proposed approach is based on a one-class classifier that employs one-class KNN and chooses the best features using various feature selection techniques including PCA and XGBoost. The proposed approach is implemented by using various datasets gathered from different IOT devices. The experimental findings demonstrate enhanced performance on the collected datasets. The proposed method exhibits excellent accuracy, which is 99% and is successful in identifying IoT botnets early on. This will lead to a decrease in the effects of impending DDoS attacks.