To evaluate proportion and predictors of poor response in infants and appraise booster seroprotection, we surveyed 2047 infants in Shanghai and detected antibody to hepatitis B (HB) surface antigen (anti-HBs). Poor responders were randomized into 2 groups, given booster with 5 µg and 10 µg hepatitis B vaccine (HepB), respectively. Proportion of infants with titer <10 mIU/mL and 10 to 99 mIU/mL was 1.86% and 15.14%, respectively. Multivariate logistic regression suggested infants of male, aged 13 to 18 months, premature, administered with 5 µg HepB or mother positive for HB surface antigen (HBsAg) and HBe antigen (HBeAg) would more likely develop worse response. Difference of geometric mean concentration between the first and full booster was not statistically significant both for 5 µg and for 10 µg HepB groups. The seroprotective rate were higher for infants with 10 µg HepB than those with 5 µg HepB (P > .05). Therefore, it is concluded that booster for poor vaccinees with 10 µg HepB could achieve satisfactory seroprotection.