Automated region classification of Moroccan vehicles is critical for emerging smart city applications, including traffic management and law enforcement. Existing systems struggle with Arabic script and complex plate layouts, prompting our focus on numeric region codes (e.g., ‘44’ for Tetouan) located on Moroccan plates. We collected a dataset of 2,817 license plates and fine-tuned YOLOv11 for detection, achieving 98% . After isolating numeric regions, we applied EasyOCR for digit recognition, reaching 93% accuracy with preprocessing techniques like resizing and thresholding. Preprocessing mitigated motion blur and lighting issues, improving accuracy by 25% compared to raw images. By prioritizing numeric code extraction, our approach sidesteps the challenges of Arabic script and enables lightweight, efficient vehicle classification with computational efficiency (<50 ms/image). This system provides a practical, low-cost solution for region-based vehicle tracking without the need for expensive infrastructure upgrades.