The study was designed to compare the independent predictors of one-year mortality after ischemic stroke in patients with and without diabetes. We studied 667 patients with ischemic stroke, including 166 diabetic patients and 501 patients without diabetes. The diagnosis of ischemic stroke was confirmed in each case by computed tomography (CT) of the head. We collected data on patients' sex, age, and on the presence of major risk factors for stroke. The neurological deficit, including the disorders of consciousness, was assessed by Scandinavian Stroke Scale on admission. The body temperature was measured on admission and then every four hours during the first day of hospitalization. The outcome was measured by 30-day and one-year all cause mortality. The neurological deficit and the percentage of patients with large ischemic lesions in CT scanning were similar in patients with and without diabetes. Lacunar infarcts were more frequently seen in diabetics (30.1 vs 22.5%; p < 0.05). Diabetic and non-diabetic patients had similar 30-day and one-year mortality (16.3 vs. 17.6% and 41.0 vs. 35.1%, respectively; p = ns, chi2 test). One-year mortality among patients with non-lacunar infarcts was greater in diabetics (51.7%) than in non-diabetics (35.3%; p < 0.01). The multinomial logistic regression analysis revealed that older age and greater neurological deficit were independent predictors of death within one year after stroke in diabetic patients, while older age, greater neurological deficit and increased body temperature during the first day of hospitalization independently predicted one-year mortality in non-diabetic patients.