The assessment of fetal growth is usually based on classifying the estimated fetal weight (EFW) into a reference curve, which may over or underestimate growth restriction. The main limitation of this technique has been the undeterminable bias of these birthweights as reference for the normal population, since premature birth itself is related to pathological states that may affect the intrauterine growth. We propose a new method based on longitudinal approach to examine the rate of EFW increment along gestation. A one year period was considered in this study. The inclusion criteria were: 1) more than 3 ultrasound scans; 2) normal pregnancies delivered at term. EFW were calculated using the equation published by Hadlock et al. Ultrasound examinations were performed by experienced physicians (Voluson 730-Pro). Equation analysis was based on linear model, after logarithmic transformation of the gestational age (GA) and EFW. Normal values for the constant (C) and slope (S) with its 99% lowest confidence intervals (LCI) were calculated. 355 patients were included in the study. Linear model curve fit was R2 = 0.98. Normal parameters were: C = − 4.147 (SE = 0.066); S = 3.375 (SE = 0.019); LCI = 3.278. The method was tested on a second set of patients for validation: curve parameters were similar to the first set (R2 = 0.98) for those pregnancies within 10th and 90th percentiles (C = − 3.935, SE = 0.049; S = 3.307, SE = 0.014; LCI99% = 3.278), but different for those < 10th percentile (C = − 3.69, SE = 0.145; S = 3.164, SE = 0.044; LCI99% = 3.078). A subgroup of patients within 10th and 90th percentiles showed similar growth parameters than those < 10th percentile. This method may provide a better way to distinguish patients with intrauterine growth restriction from those constitutionally small for gestational age, and could better define fetuses at risk for intrauterine growth restriction regardless of their percentile classification.