Objective: Our purpose was to compare the accuracy of ultrasonographic fetal weight estimation in pregnant diabetic women with that of matched nondiabetic controls. Study design: We performed a case-control study of pregnant patients who underwent ultrasonographic fetal weight estimation within 3 days of delivery. The study group consisted of pregnant diabetic women and nondiabetic controls matched for maternal body mass index and neonatal birth weight. Fetal weight estimates were calculated with use of Hadlock's and Shepard's formulas. The difference between ultrasonographic fetal weight estimation and actual birth weight (absolute percent error) was analyzed with respect to maternal diabetic status and actual birth weight. Results: A total of 450 patients were studied (225 patients in each group). The mean (±SD) gestational age at delivery was 39.0 ± 1.5 weeks versus 39.9 ± 1.7 weeks for the diabetic and nondiabetic patients, respectively. There was no statistically significant difference between the two groups with respect to the mean (±SD) time interval between the ultrasonographic examination and delivery (0.9 ± 1.8 days vs 0.8 ± 2.1 days) or the mean (±SD) absolute percent error (9.0% ± 7.1% vs 8.4% ± 6.3%). The mean (±SD) absolute percent error of fetal weight estimates among subjects with macrosomic fetuses (birth weight ≥4500 gm) was significantly greater than that observed in fetuses with birth weights <4500 gm (12.6% ± 8.4% vs 8.4% ± 6.5, p = 0.001). This difference was observed irrespective of maternal diabetic status. Conclusion: When matched for maternal body mass index and birth weight, the accuracy of ultrasonographic fetal weight estimation was similar among diabetic and nondiabetic women. Birth weights ≥4500 gm rather than maternal diabetes seem to be associated with less accurate ultrasonographic fetal weight estimates.