AIM : The aims of our study were to develop new maturity scores for dental age estimation in South African black children according to the Willems method, which was developed based on Belgian Caucasian (BC) reference data (Willems et al. J Forensic Sci 46(4):893–895, 2001), and to compare age prediction performance of both methods. SUBJECTS AND METHODS : A total of 986 panoramic radiographs of healthy South African black (SAB) children (493 males and 493 females) in the age range of 4.14 to 14.99 years (mean age 10.06 years) were selected for obtaining developmental staging scores (according to Demirjian et al. Hum Biol 45(2):211–227, 1973). Willems BC methodology was applied to develop new country-specific maturity scores (Willems SAB). Age prediction performance of Willems BC and Willems SAB was compared. RESULTS : On average, Willems BC renders acceptable results with an overestimation of chronological age of 0.06 years (SD 0.88 years) in SAB children. Compared to Willems SAB, the overall mean absolute error was slightly higher with Willems BC (0.62 and 0.68 years, respectively), but this was not significant in males. Also, the root mean squared error was marginally higher in Willems BC. CONCLUSION : The new age prediction method developed in South African black children was found to be better compared to Willems BC, although the difference seems to be small and clinically not relevant, especially in males.