Detecting diseases early is crucial for maintaining crop health and maximizing yield of Rice Plants. Traditional manual inspection methods required not only time but also susceptible to human error, which limits the scalability and efficiency of crop monitoring. In this study, we present a lightweight deep learning model, YOLOv11 designed for precise and real-time detection of rice plant diseases. To train and evaluate the model, we utilised a custom dataset of 4517 annotated images spanning five categories-leaf blast, brown spot, bacterial leaf blight, tungro, and healthy. The test results show that YOLOv11 has a precision of 93.5%, a mean Average Precision (mAP) of 83.4%, and F1-score of 89.9%, a recall of 91.7%, which is better than other new models, context of both speed and accuracy of detection. The model’s low complexity and high inference efficiency make it ideal for deployment on mobile and edge devices. Comparing this approach with other novel models confirms the effectiveness of integrating high-quality datasets with modern, lightweight architectures for smart agriculture applications and real-time disease management.