We compared the efficacy and safety of cilostazol and aspirin in 2,672 Japanese patients with non-cardioembolic ischemic stroke. The patients were randomized to be allocated either on cilostazol (200 mg/day) group or aspirin (81 mg/day) group, and were followed up for one to five years (average 29 months). The primary endpoint was any stroke, and safety endpoint was hemorrhagic stroke or hemorrhage requiring hospitalization. Annual incidence of stroke was significantly lower in the cilostazol group (2.76%) than in the aspirin group (3.71%) (relative risk reduction [RRR] 25.7%, p=0.0357) and annual incidence of hemorrhagic stroke or hemorrhage requiring hospitalization was 0.77% in the cilostazol group and 1.77% in the aspirin group (RRR 54.2%, p=0.0004). The sub-analyses between subtypes of ischemic stroke showed that annual incidence of hemorrhagic stroke was much lower in the cilostazol group (0.36%) than in the aspirin group (1.20%) among patients with lacunar stroke (p=0.003). The results suggest that cilostazol has a favorable risk-benefit profile alternative to aspirin for secondary stroke prevention in patients with non-cardioembolic ischemic stroke, particularly in patients with lacunar stroke, who are at high risk of hemorrhagic stroke.