Objective To determine the optimal cut point on the NIH Stroke Scale (NIHSS) for predicting poor 90-day clinical outcome in patients with supratentorial and infratentorial acute ischemic stroke (AIS). Methods Data are from participants of the alteplase-dose arm of the randomized controlled trial, Enhanced Control of Hypertension and Thrombolysis Stroke Study (ENCHANTED). Associations between baseline characteristics of clinically defined supratentorial and infratentorial AIS patients and poor functional outcome, defined by scores 3-6 on the modified Rankin Scale, were evaluated in logistic regression models, with area under the curve (AUC) receiver operating characteristics defining the optimal NIHSS predictor cut point. Results Patients with infratentorial AIS (n = 289) had lower baseline NIHSS scores than those with supratentorial AIS (n = 2,613) (median 7 vs 9; p < 0.001). NIHSS cut points for poor outcome were 10 (AUC 76, sensitivity 65%, specificity 73%) and 6 (AUC 69, sensitivity 72%, specificity 56%) in supratentorial and infratentorial AIS, respectively. There was no significant difference in functional outcome or symptomatic intracranial hemorrhage between AIS types. Conclusions In thrombolysis-eligible AIS patients, the NIHSS may underestimate clinical severity for infratentorial compared to supratentorial lesions for a similar prognosis for recovery. Because thrombolysis treatment has low effect on stroke outcome in patients with infratentorial AIS when baseline NIHSS score is more than 6, additional treatment such as endovascular treatment should be considered to improve stroke outcome.