to determine the concurrent validity of the System for Observing Fitness Instruction Time (SOFIT) to evaluate students’ physical activity levels during physical education classes in three public schools from Bogotá, Colombia. A cross – sectional study was conducted between October 2014 and May 2015. The sample collected 48 students (25 girls and 23 boys) aged 5 and 17 years old in 1st to 9th grade. Students were selected accordingly to SOFIT protocol. The number of minutes and percentage lesson time sport in sedentary behavior, moderate physical activity (PA), vigorous PA and moderate to vigorous PA (MVPA) were evaluated by SOFIT and then validated using accelerometry data as standard. The validation was calculated through the mean absolute difference, mean signed difference, linear regression and fixed effects model methods. The correlation coefficient between SOFIT and accelerometry was high for moderate PA (Rho=, 968; p=, 000), vigorous PA (Rho=, 937; p=, 000) and MVPA (Rho=, 962 ; p=, 000). The positive correlation was also found on the fixed effects model (moderate PA ß1=, 95 (p=, 000), vigorous PA ß1=, 94 (p=, 000) and MVPA ß1=, 95 (p=, 000)). The results showed no meaningful differences between the two methods used to measure the PA levels. Sedentary behavior was positively correlated with Spearman (Rho=, 965; p=, 000), but showed discrepancy with the fixed effects model. SOFIT is an appropriate, low cost and valid method to evaluate students’ PA levels during physical education classes within the school context. These findings should encourage schools to conduct research to assess PA levels and incorporate strategies to promote PA among students.