Abstract: In view of the problem that there is no standard diagnosis for early stage keratoconus disease,at the same time to assist the special examiner and ophthalmologist to make the early diagnosis effectively,the advantages and disadvantages of each testing instrument were analyzed.In order to construct an assistant system for early diagnosis of keratoconus,a deep learning technique was applied in corneal OCT examination.The system used improved VGG-16 to realize the recognition accuracy of about 68% keratoconus keratopathy,and the clinical results showed that the system can help doctors to give diagnosis confidence to a certain extent.At the same time,the physician's re-marking of OCT can help train the system for more accurate judgment.