Midazolam has been widely studied for preventing emergence agitation. The authors previously reported that in children with sevoflurane anesthesia, intravenous administration of midazolam (0.05 mg/kg) before the end of surgery reduced the incidence of emergence agitation but prolonged the emergence time. This study was designed to test the hypothesis that a lower midazolam dose could suppress emergence agitation with minimal disturbance of the emergence time in children with sevoflurane anesthesia. In this randomized, double-blind, placebo-controlled trial, 90 children (1 to 13 yr of age) having strabismus surgery were randomized to 1:1:1 to receive 0.03 mg/kg of midazolam, 0.05 mg/kg of midazolam, or saline just before the end of surgery. The primary outcome, the incidence of emergence agitation, was evaluated by using the pediatric anesthesia emergence delirium scale and the four-point agitation scale. The secondary outcome was time to emergence, defined as the time from sevoflurane discontinuation to the time to extubation. The incidence of emergence agitation was lower in patients given 0.03 mg/kg of midazolam (5 of 30, 16.7%) and patients given 0.05 mg/kg of midazolam (5 of 30, 16.7%) compared with that in patients given saline (13/of 30, 43.3%; P = 0.036 each). The emergence time was longer in patients given 0.05 mg/kg of midazolam (17.1 ± 3.4 min, mean ± SD) compared with that in patients given 0.03 mg/kg of midazolam (14.1 ± 3.6 min; P = 0.0009) or saline (12.8 ± 4.1 min; P = 0.0003). Intravenous administration of 0.03 mg/kg of midazolam just before the end of surgery reduces emergence agitation without delaying the emergence time in children having strabismus surgery with sevoflurane anesthesia.