Résumé : Objective: To prospectively compare magnetic resonance (MR) estimation of fetal weight (MR-EFW) performed at third trimester with ultrasound (US) estimation of fetal weight (US-EFW) and actual birth weight, and to evaluate factors influencing fetal growth rate near term. Methods: US-EFW and MR-EFW were calculated at a median of 33.0 and 37.7 weeks of gestation in 37 fetuses and plotted on curve centiles to predict birth weights at 39.3 weeks of gestation. The median absolute relative errors for predicted US-EFW and MR-EFW were calculated. Regression analysis was used to investigate the effect of different variables on fetal growth rate at 35.2 weeks of gestation. Results: The relative error of actual birth weight as predicted by US at 33.0 weeks was significantly higher compared with MR (7.33 vs. 4.11%; p = 0.001). This was also the case for fetal weight predicted by US at 37.7 weeks as compared with MR (6.63 vs. 2.60%; p < 0.01). Fetal growth rate was significantly and independently positively associated with the mother's weight and with gestational age at estimation (p < 0.05 for both variables). Conclusion: Fetal weight estimates predicted using MR at third trimester are better than those given by prenatal US. Fetal growth rate depends on fetal and maternal characteristics.