Global agricultureAgriculture is seriously threatened by plant diseases, which lower agricultural production and cause financial losses. Deep learningDeep learning models have made significant progress in the past few years in automating the identification of leaf diseases, allowing for early intervention and customized therapies. This study focuses on the use of three cutting-edge deep learningDeep learning models for the detection and classification of different types of illnesses on grape and potato leaves: YOLOv5YOLOv5, YOLOYOLO-NAS, and YOLOv8YOLOv8. Our research focused on the identification of specific leaf diseases, including leaf blightLeaf blight, ESCAESCA, and black rotBlack rot in both grapevine and potato leaves. The primary objective was to enhance the speed and accuracy of these algorithms for real-time detection and classification. The research results revealed a remarkable mean Average Precision (mAP) value of 98.6% for the YOLOv8YOLOv8 model. This finding suggests that, among the three models we trained, YOLOv8YOLOv8 demonstrated the highest level of accuracy and efficiency in detecting and classifying leaf diseases.