The aim of the study. To compare different methods for assessing comorbidity in terms of its long-term predictive value after myocardial infarction (MI).Materials and methods. The analysis included 1176 patients with MI who were consecutively admitted to the hospital. The incidence of STsegment elevation MI was 60%; every second patient underwent endovascular intervention. All patients underwent an analysis of the severity of comorbidity according to the CIRS system (Cumulative lllness Rating Scale), according to the CCI (the Charlson’s comorbidity index), the CDS scale of chronic diseases (Chronic Disease Score), as well as according to their own model ‘K9’ (patent RU2734993C1 dated 10.27.2020) based on the summation of nine diseases: type 2 diabetes mellitus, chronic kidney disease, atrial fibrillation, anemia, stroke, arterial hypertension, obesity, peripheral atherosclerosis, thrombocytopenia.Results. Long-term mortality was 12.1 %. In Cox regression analysis of long-term survival after MI, the K9 model showed the best operational characteristics with a p < 0.00001 level. In multivariate analysis, when comorbidity data were added to GRACE, an increase in the χ2 value for GARCE + CCI and GRACE + K9 to 102.5 and 99.3, respectively, and the values of the area under the ROC curve to 0.78 (0.74–0, 82) and 0.77 (0.72–0.81), respectively. Regardless of the initial level of risk assessed by the GRACE scale, severe comorbidity (four or more diseases according to the K9) significantly increased the relative risk of mortality. In patients with severe comorbidity, the predictive value of the GRACE scale was the lowest.Conclusions. Among the analyzed methods of assessing comorbidity, only CCI and its own K9 scale have an acceptable predictive value, allowing better adaptation of the GRACE scale for stratification of the long-term risk of death after MI. At the same time ‘K9’, based on the summation of nine previously described diseases, is much more convenient than CCI in practical application