In leaf disease detection, the need for efficient and accurate methods keeps growing. Traditional techniques like visual inspection or microscopic analysis are often slow. They also depend heavily on human judgment, which makes them prone to errors. Recent advances in computer vision and image-based detection offer a promising alternative. These technologies reduce subjectivity and speed up the diagnostic process. Among them, object detection algorithms—especially YOLO—stand out. They have shown strong results in identifying plant diseases, even in datasets with complex visual patterns. This study evaluates the YOLOv9 algorithm for detecting diseases in plant leaves. We trained and tested our models using the PlantDoc dataset, which includes a wide variety of annotated leaf images.YOLOv9 outperforms other versions. It achieved the highest mAP@0.5 score (0.5410), compared to YOLOv5s (0.497) and YOLOv8s (0.496). It also reached the best mAP@0.5:0.95 score (0.3880), which shows better accuracy in both object localization and classification across different IoU thresholds. These results suggest that YOLOv9 is more precise and robust for leaf disease detection tasks.