Introduction: The estimation of bone mineral density (BMD) is closely related to the health status of schoolchildren) at early ages, the estimate value allows to identifying variables that can directly affect health during aging, and also generates a way of low cost, easy access and development. The estimate value is obtained by anthropometric techniques compared to the DEXA study, which is listed as a high-cost gold test for the evaluation of BMD; therefore, this research had the purpose to determine the predictive variables of (BMD). Materials and methods: for the research there was a random sample of 207 schoolchildren between 8 and 16 years old to the city of Cali, of those 207 schoolchildren 142 belong a public schools and 65 belongs a private schools, 121 are men and 86 are women. For this sample, clinical variables such as (fractures, medication consumption, supplement consumption), level of physical activity with the PAQ-C questionnaire, level of physical activity according to age, body mass index (BMI), weight, vertical height (AV) sitting height (AS) right forearm length (LAD), left forearm length (LAI), femur diameter (DF) lower limb length were analyzed and finally relationships and correlations are made between BMD and study variables. Results: the population evaluated belongs mostly to the low socioeconomic stratum with 63.8%, the middle stratum with 30.9% and the high stratum with 5.3%. These schoolchildren are in grade 1 to grade 11. 79.7% have had no fractures, 10.6% consume medications and 84.5% do not consume nutritional supplements. They present a low level of physical activity for both sexes, but women do less physical activity. The population presents a normal body mass index; there is a strong and positive correlation between BMD and anthropometric variables except for femur diameter. The predictive model for schoolchildren is determined as follows: 0.199 + 0.028 * (age) + 0.001 * (vertical height) + 0.004 * (sitting height) -0.012 (- growth peak) + 0.022 * (forearm length right) + 0.13 * (left forearm length) with an overall adjustment of 66.3%, F = 65.611, sig = 0.000.