In modern agriculture, accurately anticipating and managing cucumber diseases has a major impact on crop output and sustainability. This research presents a deep learning algorithm based advanced methodology for the prediction of cucumber disease. We illustrate the effectiveness of deep learning in identifying different cucumber diseases namely Alternaria leaf spot, Cercospora leaf Spot, Powdery mildew, Downy mildew, and viral diseases by applying the most advanced COCO and YOLO models, which allows for proactive disease control. Together with these methods, we make use of Roboflow’s simplified features for effective model training and data annotation. On running COCO model over augmented preprocessed data, a mAP of 60% and precision of 78.7% was obtained. Further, the model was deployed on the Roboflow platform.