<p>The attack surface has expanded exponentially with the proliferation of Internet of Things (IoT) devices, as interconnected systems and their underlying network infrastructures have become increasingly vulnerable. A compromise of even a single IoT component can jeopardize the security of the entire network, potentially leading to data breaches and enabling the compromised devices to launch further attacks. Addressing these challenges requires robust and intelligent defense mechanisms. Machine Learning (ML) provides a promising solution to detect and mitigate such threats by learning complex attack patterns and adapting to evolving behaviors. The objective of this study is to develop and evaluate binary classification ML models for detecting botnet attacks in IoT environments. To achieve this, several algorithms Decision Tree (DT), Logistic Regression (LR), Gradient Boosting (GB), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) were applied to the BoT-IoT dataset. To overcome issues related to class imbalance and suboptimal performance, the models were enhanced through the integration of the Synthetic Minority Oversampling Technique (SMOTE) and hyperparameter optimization using GridSearch Cross-Validation (GridSearchCV). The experimental results demonstrate that the DT and GB models achieved the highest performance among all evaluated models, confirming their suitability for IoT botnet detection tasks. After applying SMOTE and hyperparameter tuning with GridSearchCV, the overall model performance improved significantly. The Decision Tree and Gradient Boosting models achieved 100% accuracy, while Logistic Regression closely followed with 99.9959%, confirming the effectiveness of the optimization process.</p>