Skeletal age (SA) is considered the gold standard to assess the degree of maturation and has been widely used in sports, education and public health areas; however, it requires sophisticated equipment and well-trained technicians. Therefore, it is important to develop non-invasive methods for its evaluation. The aim was to develop an equation to predict SA using the percentage of adult height. SA was measured by Tanner-Whitehouse-3 method, and the percentage of adult height was estimated by two methodologies: Tanner-Whitehouse-3 method (P-TW3) and Khamis-Roche method (P-KR) using 839 schoolchildren of both sexes. Linear regression was used for predicting SA from P-TW3; then P-TW3 was replaced in the equation for P-KR value. Bland–Altman graphs, interclass correlation coefficient and Kappa index were used as validation tests. Model showed a SA predictive capacity of 93.2% in boys and 96.8% in girls. The average differences between SA measured and SA predicted by P-TW3 was 0.0504 (± 0.664) in boys and 0.0144 (± 0.435) in girls (P = 0.229 and 0.667, respectively). When P-TW3 was replaced for P-KR value in the equation, the average differences were − 0.0532 in boys and 0.0850 in girls (P = 0.509 and 0.167 respectively). The present model, based on the percentage of adult height, showed an adequate estimation of SA in children and adolescents and it can be used in the absence of bone X-ray equipment, in healthy boys aged 9 to 15 and girls 8 to 13.