BACKGROUND: The relationship between diabetes control status and long-term prognosis after stroke incidence remains unclear. This study aimed to investigate the effect of diabetes status at admission on long-term survival in patients with first-ever stroke. METHODS: A retrospective cohort study was conducted based on the Shiga Stroke and Heart Attack Registry in Japan. Patients were classified according to their diabetes status and glycated hemoglobin (HbA1c) value at hospital admission into the following: (1) free of diabetes (no history of diabetes and HbA1c <6.5%); (2) good control (history of diabetes and HbA1c <7%; free of history and 6.5% ≤HbA1c <7%); and (3) poor control (with or without a history of diabetes and HbA1c ≥7%). Multivariable Cox regression models were used to evaluate the association between diabetes status and long-term survival from stroke onset. Additionally, we also evaluated the association between diabetes status and conditional survival, beginning 29 days after stroke onset. RESULTS: A total of 6,331 first-ever stroke patients were eligible for this study. Among study patients, the mean (±SD) age was 72.85 ± 13.19 years, and the mean (±SD) follow-up year was 2.76 ± 1.66 years; additionally, 42.09% of patients were women. Among patients with all strokes, considering the free-of-diabetes group as the reference group, the adjusted hazard ratio (95% confidence interval) for mortality was 1.26 (1.10, 1.44) in the good control group and 1.22 (1.05, 1.41) in the poor control group. Among patients with ischemic stroke, the adjusted hazard ratio was 1.24 (1.06, 1.46) in good control group and 1.27 (1.08, 1.50) in poor control group. After excluding patients who died within 28 days, the adjusted hazard ratio for conditional mortality in the poor control group was 1.31 (1.12, 1.54) among all stroke patients and 1.29 (1.08, 1.54) among ischemic stroke patients. No significant associations were observed between diabetic status and long-term mortality in intracerebral hemorrhage patients. CONCLUSIONS: The findings suggest that first-ever stroke patients with diabetes exhibited a higher risk of all-cause mortality than those without diabetes, particularly in the overall stroke and ischemic stroke populations. Additionally, in stroke populations after 28 days of onset, high risk of long-term mortality was stated in stroke patients with poor HbA1c control.