Identifying and preventing botnet attacks has become increasingly difficult due to the explosive growth of IoT devices. This research suggests a useful method for detecting IoT botnet attacks that uses a Random Forest classifier to examine network traffic data and spot suspicious activity. EDA is used to analyze the dataset's structure, identify missing values, and evaluate the distribution of classes. Categorical features are encoded with labels to make them compatible with machine learning algorithms. A Random Forest classifier is selected to its capacity to effectively handle skewed distributions and dimensional data, taking into account the dataset's intrinsic class imbalance. Using the classifier's integrated ranking mechanism, feature importance analysis is carried out, choosing only the most pertinent features to improve mode ln performance. The data is then classified into training and testing sets, with the most important features being used to train the model. Accuracy, classification reports, and F1-score are used to assess the system, showing that the Random Forest classifier accurately and efficiently detects IoT botnet attacks. This study emphasizes how important feature selection, data pretreatment, and machine learning models are to bolstering IoT network cybersecurity defenses.