Potholes on roads pose significant risks to safety and infrastructure, necessitating effective detection and maintenance strategies. This study employs YOLOv8, a state-of-the-art deep learning model, to develop a reliable and efficient pothole detection system. YOLOv8 is distinguished by its advanced architecture, anchor-free detection methodology, which enhance its precision and speed. The model achieved high detection accuracy, as evidenced by its performance on precision-recall curves, F1 scores, and confusion matrices. The results demonstrate YOLOv8’s ability to accurately identify and localize potholes under varying road and environmental conditions in real time. This work highlights YOLOv8’s potential to improve road safety through automated infrastructure monitoring, offering a practical solution for urban maintenance systems.