A chronic disease that is becoming more common worldwide, diabetes mellitus requires early diagnosis and treatment to prevent complications. In healthcare, artificial intelligence (AI) has become a game-changing technology, especially when it comes to identifying complicated conditions like diabetes. The use of AI-driven data mining approaches to increase the precision, effectiveness, and dependability of diabetes diagnosis is investigated in this work. Medical imaging has benefited greatly from deep learning (DL), a kind of artificial intelligence (AI) built on deep neural networks, especially in the areas of pattern identification and picture categorization. A class of eye disorders known as diabetics harms the optic nerve. Good eyesight depends on the optic nerve, which transmits visual information through the eye to your brain. High eye pressure is frequently linked to harm done to the optic nerve. However, even with normal ocular pressure, diabetes can develop. The optic nerve, which transmits images to the brain, may sustain damage from elevated intraocular pressure. Within a few years, diabetes can result in irreversible vision loss or possibly complete blindness if the damage gets severe. Regular eye exams are essential for the early diagnosis and treatment of conditions like diabetes. You can prevent serious visual damage by receiving therapy early. We used CNN’s Image Net-trained models to perform a diabetic evaluation with fundus photos in order to solve this issue. This study highlights how AI and data mining might transform diabetes care by tackling issues including unbalanced datasets, computational cost, and model interpretability. The results highlight how important it is to incorporate AI-based tools throughout clinical practice in order to enhance the provision of healthcare and supplement human expertise.