Agriculture is one of the most functional ancient cultivation forms that is now updating itself with many modern techniques. Its contribution to the economy might be a bit low in some countries but its impact on society always remains at the top, it is impossible to imagine a day on worldwide scale with no food or agricultural products. By saying the importance of agriculture and knowing that plants are the atomic units in agriculture, the discussed importance is reflected same as the importance of detection of disease in plants, with the help of state of arts technology like deep learning. Thus, this paper proposed a multi-staged deep learning approach for detection of disease in plants with the help of methods like diffusion model like Imagen and detection model like You Only Look Once (YOLO-V8). The proposed model utilizes the self-collected images of different plant parts which are further augmented and increased with the help of imagen and then that dataset is further processed and YOLO-V8 was assigned responsibility to detect the diseased parts in image. The proposed has achieved training accuracy of 96.8% and testing accuracy of 95.9%.