Estimated fetal weight (EFW) is an important metric at delivery as neonates with abnormal birthweight and their mothers are at higher risk of birth complications. Data regarding optimal EFW assessment in gravidas with obesity is inconsistent, and with the increasing incidence of obesity, clarification of this question is crucial. We aimed to compare accuracy of ultrasound (US)-derived EFW and clinical assessments of EFW in predicting neonatal birthweight among gravidas with obesity. This prospective cohort study enrolled gravidas with obesity and a singleton pregnancy admitted for delivery at term. EFW was determined using either US biometry or clinical assessment (Leopold's maneuvers, Johnson's formula, and Insler's formula) at time of admission. Our primary outcome was accurate EFW, defined as EFW within 500 g of birthweight. Secondary outcomes included ability to predict small-for-gestational age (SGA) and large-for-gestational age (LGA) birthweights. These outcomes were compared between all EFW methods. A total of 250 gravidas with a median body mass index of 36.4 kg/m2 were enrolled. Admission US outperformed Leopold's maneuvers in obtaining accurate EFW (81.6% versus 74.5%, P = .03). When comparing all methods, Johnson's and Insler's formulae performed the worst, accurately predicting EFW in only 27.4% and 14.3% of cases, respectively. Likewise, US-derived EFW outperformed Leopold's maneuvers and fundal height in the prediction of SGA and LGA neonates. US is more accurate than clinical assessment of EFW in gravidas with obesity both for estimation of actual birthweight and prediction of abnormal birthweight. Universal late third-trimester or peripartum US for EFW should be considered in gravidas with obesity.