Background: Our purpose was to determine whether the onset-to-arrival time affects the outcome of stroke patients. Methods: We carried out a prospective multicenter study involving 1,817 patients with ischemic stroke and 1,226 with intracerebral hemorrhage who presented to hospitals within 24 h of symptom onset. The primary outcome was independent activity of daily living corresponding to a modified Rankin Scale (mRS) score â€2 at discharge approximately 3 weeks after stroke. Results: In ischemic stroke patients, the initial NIH Stroke Scale (NIHSS) score decreased as the onset-to-arrival time increased: 9 (median) in the earliest tertile group (<3 h), 5 in the second tertile group (3â8 h) and 4 in the latest tertile group (â„8 h, p < 0.001). The median mRS scores at discharge in these groups were 3, 2 and 2, respectively (p < 0.001). After adjustment for underlying features and the initial NIHSS score, the independent activity of daily living at discharge was 1.73 times more common in patients in the earliest group than in the latest group (95% CI = 1.24â2.42, p = 0.001). A similar tendency was shown in the subanalysis for large-artery atherosclerosis and cardioembolic stroke. In intracerebral hemorrhage patients, both the initial NIHSS score and the mRS score at discharge decreased as the onset-to-arrival time increased. After multivariate adjustment, the independent activity of daily living was 2.33 times (p < 0.001) and 1.69 times (p = 0.006) less common in patients in the earliest (<1.2 h) and second tertile groups (1.2â3.5 h), respectively, than in the latest tertile group (â„3.5 h). Conclusions: Early hospital arrival improved the clinical outcome in ischemic stroke patients but not in patients with intracerebral hemorrhage.