Purpose: The purpose of this study is to validate the Mortality in Emergency Department Sepsis (MEDS) score in older patients with sepsis who visited the emergency department (ED). Methods: This was an observational study. Patients 65 years of age or older with sepsis who were admitted from January, 2010, to May, 2011, in an urban ED with approximately 30,000 annual visits were eligible. Demographic, anthropometric, hemodynamic, and laboratory data were collected. MEDS and Geriatric Nutritional Risk Index (GNRI) scores were calculated as originally described. The area under the receiver operating characteristics curve (AUC) was used for discrimination of each score. Univariate and multivariate analyses were performed. The primary end-point was 28-day in-hospital mortality. Results: Of 397 patients, 11.8%(95% confidence interval, 8.7%~15.0%) died. Mortality stratified according to MEDS was as follows: 0 to 4 points, 0.0%; 5 to 7 points, 4.8%; 8 to 12 points, 11.4%; 13 to 15 points, 22.2%; and ≥ 16 points, 35.7%. Mortality stratified according to GNRI was as follows: > 98 points, 3.9%; 92 to 98 points, 9.8%; 87 to < 92 points, 6.9%; 82 to < 87 points, 20.0%; and < 82 points, 34.6%. AUCs were 0.707 and 0.734 for MEDS and GNRI, respectively. In multiple logistic regression analysis, GNRI was found to be an independent predictor of 28-day in-hospital mortality, however, MEDS was not. Conclusion: The MEDS score performed with moderate accuracy for prediction of 28-day in-hospital mortality in older ED patients with sepsis. These results suggest the need of geriatric-specific risk assessment tools in the ED.