Objective To examine the association between antenatal depression and infant development after controlling for confounding factors. Methods A hospital-based prospective cohort study (Hokkaido Study on Environment and Children's Health) was conducted between July 2002 and October 2005 in Sapporo, Japan. Of 309 mothers who delivered at Sapporo Toho Hospital during the study period and who agreed with the clinical assessment of depression, 154 mother-infant pairs were eligible for analysis. Antenatal depression was assessed between the second and third trimesters using the Edinburgh Postnatal Depression Scale (EPDS), and infant development was assessed at 6 months by the Bayley Scales of Infant Development II (BSID-II). Data on potential confounders, including socioeconomic status, birth complications, postnatal depression and child care environment, were obtained from medical records and selfadministered questionnaires. Univariable and multivariable analyses were conducted in which the EPDS score was entered as an independent variable and the BSID-II scores as a dependent variable, adjusting for confounders. Results Although the antenatal EPDS score tended to be related to the BSID-II score in the univariable analysis, this correlation was lost in the multivariable analysis. However, based on a series of linear regression analyses, antenatal depression was found to be significantly related to shorter gestational age (b = -0.25, 95 % confidence interval (CI) [-1.20, -0.17]), and shorter gestational age was significantly related to a lower BSID-II (mental development) score (b = 0.23, 95 % CI [0.00, 0.00]). Conclusions Gestational age is an important confounder in the association between maternal antenatal depression and infant development. A delay in infant development may be related to a shorter gestational period caused by maternal depression during pregnancy.