Objective To compare the diagnostic value of estimated glomerular filtration rate (eGFR) by various equations of Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) in evaluation of the prognosis in the elderly patients with chronic kidney disease (CKD), and analyze the risk factors for endpoint events. Methods A total of 682 old people, aged over 60 years, who had taken physical examination for 4 consecutive years in our health center during March 2015 to December 2018 were enrolled in this study. The eGFR was evaluated using CKD-EPI equations (CKD-EPICr, CKD-EPICys and CKD-EPICrCys) based on creatinine (Cr), cystatin C (Cys) and combination of them, respectively, and the obtained rate was evaluated for diagnostic value of endpoint events (all-cause death, cardiovascular events, acute kidney injury and rapid kidney function decline). The risk factors for these events were also analyzed. SPSS statistics 23.0 was used to analyze the data. According to the data types, Student′s t test or Chi-square test was used for comparison among groups. Receiver operating characteristic (ROC) curve was used to evaluate the diagnostic value, and logistic regression analysis was employed to analyze the risk factors affecting the prognosis. Results ROC curve analysis showed that the eGFR from CKD-EPI_cys (AUC=0.692, P<0.001) and CKD-EPICrCys equations (AUC=0.647, P＜0.001) had diagnostic value for the occurrence of endpoint events, with that from the former equation higher. The eGFR from the above 2 equations also identified same independent risk factors for the occurrence of endpoint events, that is, ratio of urinary albumin to creatinine (UACR, OR=2.263, 95%CI 1.359-3.771), hypertension (OR=1.679,5%CI 1.143-2.467), anemia (OR=1.959,5%CI 1.245-3.084) and hospitalization frequency (OR=1.471,5%CI 1.321-1.637). Conclusion CKD-EPI_cys equation shows the greatest advantage in the diagnostic value of prognosis evaluation in the elderly with CKD. UACR, hypertension, anemia and hospitalization frequency are independent risk factors for endpoint events in them.