Breastfeeding self‐efficacy is a relevant predictor of exclusive breastfeeding during the first six months of life. The Prenatal Breastfeeding Self‐Efficacy Scale (PBSES) is a 20‐item tool developed to determine breastfeeding self‐efficacy during pregnancy. Our study aimed to assess the structural validity and psychometric characteristics of the PBSES and to explore item reduction according to the statistical criteria for parsimony and incremental validity. In this study, conducted in six hospitals in eastern Spain, we recruited 1183 women with healthy, full‐term, single‐birth newborns. Data on sociodemographic, breastfeeding‐related variables, and the PBSES, were obtained from self‐administered questionnaires during the third trimester of pregnancy, at postpartum discharge, and 5 months postpartum, and from a phone survey 12 months postpartum. Item reduction was conducted after revising the PBSES item floor and ceiling effects, interitem correlations, and item–score relationships with breastfeeding‐status variables during follow‐up. The factorial structure of the short form of the PBSES (PBSES‐SF) was tested using both exploratory and confirmatory approaches. After item reduction, the confirmatory factor analysis of the 12 remaining items of the PBSES‐SF revealed adequate fit statistics for a three‐factor structure and a second‐order factor. Internal consistency was measured using the Cronbach's α coefficient of the PBSES‐SF (0.86). We provided evidence on the discriminant validity of the PBSES‐SF by comparing its scores between known groups, convergent validity by examining its correlations with other variables, and predictive validity by assessing the association of PBSES‐SF scores with breastfeeding behavior at critical points in time during the first postpartum year.