To evaluate the performance and the utility of using birthweight-adjusted scores of the Ballard method of estimating gestational age in a Zimbabwean population. A validation study. Harare Maternity Hospital, from October to December 1999. Three hundred and sixty four African newborn infants, with a known last menstrual period (LMP), within 56 hours of life. Ballard scores obtained by examining the newly born infants compared to gestational age calculated from the last menstrual period. The Ballard method was a good predictor of gestational age, useful in differentiating term from preterm infants in the Zimbabwean population. There was a strong correlation between total neurological criteria (Pearson coefficient = 0.79), total physical criteria (Pearson coefficient = 0.77), total scores (Pearson coefficient = 0.81), with the gestational age calculated from the last menstrual period. The error of prediction of one single observation was 1.89 weeks. Our regression line for predicting gestational age was Y(LMP gestational age) = 24.493 + 0.420*score. Addition of birthweight to the linear regression model improved estimation of gestational age; Y(LMP gestational age)= 24.002 + 0.292*score + 0.0016*grams. The variance explained r2 was 0.63 and improved to 0.67 after addition of birthweight. The Ballard method can be used to differentiate pre-term from term infants at birth in a Zimbabwean population. The introduction of birthweight into the maturity scale improves assessment of gestational age and corrects error caused by low birthweight. We, therefore, recommend the use of our birthweight-adjusted Ballard maturity scales for routine clinical practice.