Diabetic retinopathy, which is the leading complication of diabetes mellitus, is a common ophthaliological disease for which the classical markers of apoptosis - matrix metalloproteiases - are insufficiently studied and there are no prognostic models. The aim of the study - was to develop a regression model based on indicators of the content of matrix metalloproteinases in blood plasma to predict diabetic retinopathy in old age. Material and Methods: The results of this study are based on the analysis of 63 patients with diabetic retinopathy aged 60 - 74 years, who made up the main group. The control group included 38 patients 60-74 years old without diabetic retinopathy. In both groups, a study of the content of matrix metalloproteinases in blood plasma was carried out by solid-phase enzyme-linked immunosorbent assay. A prognostic model has been developed by the regressive method Results of the study: Among elderly patients with diabetic retinopathy, there was a statistically significant increase in comparison with age control of matrix metalloproteinase - 9. Moreover, the greatest increase in the blood plasma of patients with diabetic retinopathy is characteristic of matrix metalloproteinase, which has pro-inflammatory activity. A regressive model of matrix metalloproteinases has been developed for predicting diabetic retinopathy in old age. Conclusion: The given regression model of matrix metalloproteinases makes it possible to predict the development of diabetic retinopathy in the elderly, which will make it possible to carry out adequate therapeutic and prophylactic measures in the early stages.