Introduction. Type 2 diabetes mellitus in many countries is simultaneously accompanied by an increase in the frequency of diabetic retinopathy, which is the most dangerous complication of type 2 diabetes mellitus, leading in some cases to complete loss of vision. The underestimated contribution of aging to the development of DR requires further study. The aim of the study to build mathematical models to support decision-making on screening for diabetic retinopathy in type 2 diabetes mellitus using multivariate regression analysis. Material and methods. The chronological age of the 435 patients included in the study was recorded, and then the biological age was calculated according to the Voitenko V.P. method, taking into account gender. The objective assessment was carried out using various mathematical methods – correlation, multidimensional regression, discriminant analysis, indicators of diagnostic sensitivity and specificity. Results. The values of the coefficients of multidimensional regression analysis, showing differences in the association of biological age with the content of the studied catabolic and anabolic hormones in urine in the evening hours, in two groups with the presence of dissociation between the parameters of calendar and biological age and with the absence of dissociation between calendar and biological age with ophthalmological complication caused by type 2 diabetes, demonstrate the most reliable conjugation for cortisol, melatonin and serotonin. Evaluating the practical possibility of the created discriminant models to support decision-making during screening of the discussed ophthalmological complication of type 2 diabetes, it was found that only 8,24% of the representatives with the absence of dissociation of calendar and biological age with developed as a result of type 2 diabetes in adulthood were mistakenly included in the cohort of patients with dissociation of calendar and biological age of an age with an identical complication due to type 2 diabetes in adulthood. An almost similar result was achieved when differentiating representatives into a cohort of patients with dissociation of calendar age relative to biological age and amounted to 8,85%. Conclusion. The diagnostic sensitivity of models based on biological age is high and reaches 0,92, and the diagnostic specificity is 0,88. This makes it possible to use the created diagnostic models for screening of patients with type 2 diabetes at 45-59 years of age.