AIM The pressure study was carried out to evaluate the hypotheses that ambulatory blood pressure (ABP) is superior to clinic blood pressure for ischemic stroke and that ABP is important to refine risk stratification. PATIENTS AND METHODS A total of 889 hospitalized patients who had undergone ABP monitoring were enrolled. Baseline data were compared between diabetic and nondiabetic subgroups. Area under receiver operating characteristics curve for each blood pressure (BP) component with ischemic stroke was evaluated. Each BP component was entered separately into a logistic regression basic model. The predictive performance of each model was compared using the Akaike Information Criterion and Schwartz's Bayesian Information Criterion, and a lower value indicated better performance. RESULTS The mean age of the patients was 58 years, 56.2% were men, and 63.0, 30.6, and 27.7% had hypertension, diabetes, and ischemic stroke, respectively. The diabetic subgroup had significantly higher systolic blood pressure (SBP) profiles than the nondiabetic subgroup. Compared with patients without ischemic stroke, those with ischemic stroke had significantly higher 24-h SBP and daytime-SBP, whereas no differences in clinic-SBP were observed in both subgroups. Compared with clinic-SBP, the area under receiver operating characteristics curve values for predicting ischemic stroke were significantly higher for 24-h SBP and daytime-SBP in both subgroups. The results of predictive performance showed that logistic regression models including 24-h SBP and daytime-SBP had lower Akaike Information Criterion and Schwartz's Bayesian Information Criterion values than that with clinic-SBP in both subgroups. However, no significant improvement in predictive performance was observed when the BP variable was added to the basic model. CONCLUSION Our study showed that in a hospital setting, 24-h SBP and daytime-SBP were superior to clinic-SBP in relation to ischemic stroke.