OBJECTIVE To assess and validate the best model for fetal weight estimation in the Pakistani population. METHODS The cohort included 178 women with singleton pregnancies at term (37-41 weeks) who had an ultrasound scan within 7 days of delivery. A proportionality formula was used to compensate for the fetal weight gain between the scan and the delivery. Data points from 119 pregnancies were used to derive the fetal weight formula, which was subsequently tested on the remaining 59 pregnancies. The best model for fetal weight estimation from ultrasound biometric parameters was selected. RESULTS The selected model showed an interclass correlation coefficient with birthweight of 0.89. The systematic and random errors were -10 and 250 g, respectively. This validity assessment compared favorably with the performance of the more commonly used formulae when applied to our population. CONCLUSIONS We describe a new model for fetal weight estimation in a Pakistani population which produces more valid and reliable estimates than currently used models derived from other populations.