Background: Malnutrition’s prognostic impact in patients with severe stroke requiring ICU admission is not well known. This study aimed to assess the nutritional status of severe stroke patients using the geriatric nutritional risk index (GNRI) and examine the association of GNRI with mortality in that population. Methods: We identified 1145 severe stroke patients requiring ICU admission from the Medical Information Mart for Intensive Care (MIMIC-III) database and divided them into low GNRI (≤98) or high GNRI (>98) groups. We used a propensity score matching (PSM) method to reduce confounding. Cox proportional hazards regression and restricted cubic splines were used to elucidate the association between GNRI and mortality. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated. Results: A total of 397 (35%) patients were in the low GNRI group (at risk of malnutrition). After PSM, patients in the low GNRI group still suffered higher mortality compared with the high GNRI group at 28 days (27.9 vs. 20.8%), 90 days (35.5 vs. 25.7%), and 1 year (43.4 vs. 30.9%) (p < 0.05). A low GNRI was significantly associated with an increased mortality (HR: 1.38, 95% CI 1.03–1.86 in 28 days; HR: 1.45, 95% CI 1.11–1.89 in 90 days; HR: 1.51, 95% CI 1.19–1.92 in 1 year). Sensitivity analyses yielded consistent results. Restricted cubic splines showed a progressively decreasing risk of mortality with increasing GNRI scores up to 110, approximately. Conclusion: Severe stroke patients with malnutrition experienced an increased risk of death compared to those without malnutrition. GNRI, as a simple and practical nutritional screening tool, can be used as a routine approach to the nutritional status of stroke patients.