Neonatal jaundice is a common condition affecting newborns, often requiring early detection and intervention to prevent severe complications such as kernicterus. Traditional diagnostic methods, including visual assessments and invasive blood tests, can be subjective and distressing for infants. To address these limitations, we propose a real-time, alternative non-invasive approach using Artificial Intelligence (AI) and Computer Vision (CV) techniques for real-time jaundice detection by leveraging the You Only Look Once (YOLO) object detection model. The data sets was manually annotated using Roboflow, then we develop a classification model using YOLO. Experimental results demonstrate the model’s good accuracy in distinguishing between normal and jaundiced cases. This AI-driven approach can enhance accessibility to neonatal jaundice screening, especially in resource-limited settings, reducing reliance on invasive procedures and enabling timely medical intervention.