In recent years, the proliferation of Internet of Things (IoT) devices has significantly increased the vulnerability of networked systems, particularly in fog computing environments. These environments, which operate at the edge of the network, are crucial for reducing latency and improving efficiency in data processing. However, they also present unique security challenges due to their distributed nature and limited computational resources. This research paper presents a machine learning-driven approach to attack detection specifically tailored for fog computing environments. Leveraging the dataset, our method involves comprehensive data preprocessing, dynamic sampling, and the implementation of various machine learning models, including Decision Trees, Random Forests, Support Vector Machines (SVM), and Convolutional Neural Networks (CNN). Through extensive experiments, we demonstrate the effectiveness of our approach in accurately detecting and classifying attacks, while ensuring robustness and generalization by varying the data positions in each run by HCO-DCNN. Our results indicate that the proposed machine learning-driven framework significantly enhances the security of fog computing environments, providing a reliable defense against a wide range of cyber-attacks.