Background: diabetes is the most common metabolic disorder affecting pregnancy. Its prevalence seems to be growing in parallel with the epidemics of overweight and obesity. It‘s considered a modern epidemic disease that affects about 8.3% of adults (accounts for 382 million people of the global population) and 46% of cases are estimated to be undiagnosed. Recognizing and treating diabetes or any degree of glucose intolerance in pregnancy results in lowering maternal and fetal complications. Objective: our aim is to test the hypothesis that incorporating 3D fractional thigh volume would be superior to conventional 2D biometry for predicting birth weight and macrosomia in diabetic pregnancies. Patients and Methods: the study was done on 160 pregnant women suffering from pre-gestational or gestational diabetes at El-Hussien and Sayed Galal Hospitals. All women were evaluated by full obstetrical history taking, physical examination, pelvic examination, transabdominal ultrasound, routine laboratory investigations and assessment of the route of delivery. Results: one-fifth of the patients had macrosomia, and four fifths had no macrosomia. There were no statistical significant differences between macrosomia group and no macrosomia group in maternal age, parity, body mass index, obesity, diabetes, gestational age and gestational age at delivery. The mean ± SD was lower for the 2D than the 3D projected estimate. Our study showed that there were statistical high significant regarding birth weight at delivery and difference from 2D/3D (p < 0.001). Conclusion: we suggest that 2D sonography should remain the standard of care for predicting birth weight and macrosomia in diabetic pregnancies