<i>Background:</i> Stroke prevention with oral anticoagulation is effective in patients with atrial fibrillation. An additional beneficial effect of oral anticoagulation prior to the event on stroke severity and prognosis was suggested in ischemic strokes. We tested whether this benefit is preserved after inclusion of hemorrhagic strokes and adjustment for prestroke living conditions. <i>Methods:</i> Data were used from a prospective hospital-based intervention trial evaluating quality of care and outcome in ten district hospitals. All ischemic and hemorrhagic stroke patients with atrial fibrillation were included. We analyzed separate multivariable regression models to identify factors associated with prescription of oral anticoagulation before stroke and to investigate the independent effect of anticoagulation on admission stroke severity, 3-month mortality and functional outcome. <i>Results:</i> The analysis comprised 804 (718 ischemic, 86 hemorrhagic) stroke patients admitted between July 2003 and March 2005. Males, patients aged 65–84, living independently, with diabetes, previous cerebrovascular event or additional high cardioembolic risks were more likely to receive oral anticoagulation before admission. Admission international normalized ratio (INR) between 2 and 3 (OR 0.35, 95% CI: 0.17–0.71) or higher (OR 0.32, 95% CI: 0.11–0.92) was associated with less severe strokes (including hemorrhagic strokes) compared with INR <2. Anticoagulation was associated with decreased risk of death and poor functional outcome (modified Rankin Scale >3) at 3 months (OR 0.54, 95% CI: 0.36–0.84, and OR 0.70, 95% CI: 0.47–1.06). After adjustment for stroke severity, anticoagulation had no additional effect on mortality and functional outcome. <i>Conclusions:</i> The beneficial effect of prestroke anticoagulation on stroke outcome related to the reduced stroke severity is not offset by adverse effects in hemorrhagic stokes.