Abstract Objective Both maternal prepregnancy body mass index (BMI) and gestational weight gain (GWG) influence maternal and pediatric outcomes. We sought to clarify the impact of prepregnancy BMI-specific GWG and its patterns on the risk of low birth weight (LBW) or macrosomia using data from a large nationwide study in Japan. Methods This cohort study (n = 98,052) used data from the Japan Environment and Children’s Study (JECS). The outcome variables in this study were LBW and macrosomia. We stratified the subjects into groups according to prepregnancy BMI. Results GWG from pre-pregnancy to the first trimester had a small effect on the risk of LBW and macrosomia. From the first to second trimesters, insufficient GWG was associated with the risk of LBW, and from the second trimester to delivery, a GWG of less than 2 kg was associated with the risk of LBW. These associations were commonly observed in all prepregnancy BMI categories. Irrespective of the GWG from pre-pregnancy to the first trimester, GWG from the first to second trimesters affects LBW and/or macrosomia. Irrespective of the GWG from the first to second trimesters, GWG from the second trimester to delivery affects LBW and/or macrosomia. LBW or macrosomia was associated with the prevalence of a sustained low or high BMI percentile until three years of age, respectively. Conclusions The present large national cohort study indicates that the risk of LBW or macrosomia is associated with GWG in women in Japan; the significance of this risk depends on the GWG patterns.