Internet addiction has emerged as a significant public health concern, with growing evidence linking excessive technology use to cognitive impairment and mental health disorders. This paper presents a novel EEG-based methodology for detecting Internet addiction through advanced signal processing and machine learning techniques. Our approach utilizes Laplacian energy-based feature extraction to characterize the spatial-frequency dynamics of EEG signals across all major frequency bands ( \(\delta \) , \(\theta \) , \(\alpha \) , \(\beta \) , and \(\gamma \) ). These features effectively capture the complex functional connectivity patterns in brain networks associated with addictive behaviors. Four different machine learning models are then used to classify the collected features: K-Nearest Neighbors (k-NN), Random Forest (RF), Support Vector Machines (SVM), and Logistic Regression (LR). The experimental results demonstrate exceptional performance, with Random Forest achieving 98% classification accuracy - the highest reported performance for EEG-based Internet addiction detection. This outstanding result underscores both the discriminative power of Laplacian energy features in capturing addiction-related neural patterns and the robustness of ensemble learning methods for neurological classification tasks.