Background <p>School-based intervention are needed due to the low levels of physical activity among adolescents.</p> Objective <p>To evaluate a school-based intervention’s effectiveness and potential sociodemographic moderators on physical activity (PA) among Brazilian adolescents.</p> Design <p>The <i>Movimente</i> Program was a cluster randomized controlled trial conducted at six elementary schools in southern Brazil in 2017.</p> Methods <p>Participants (<i>n</i> = 921) were adolescents from 7th -9th grade (12 to 15 years). From March to December, intervention strategies included teacher training, educational actions, and environmental improvements designed to increase PA levels. PA was self-reported, and the primary outcome was weekly minutes in total PA; secondary outcomes were weekly minutes in leisure-time PA, adherence to guidelines, and diversity (e.g., number of activities). In addition, the moderation effect of sex, school grade, socioeconomic level, weight status, and baseline PA levels was explored. Mixed-model linear regressions were used to test changes between baseline and follow-up (difference-in-differences approach).</p> Results <p>Participants were, on average, 13.1 ± 1.1 years old, and 52% were female. Pre-post changes in PA volume did not differ between intervention and control groups (β: 6.64; 95% CI: -104.60; 117.89), nor did adherence to guidelines. No significant interaction effects of sex, socioeconomic level, or weight status were found. Pre-post changes of leisure PA diversity favored the intervention group among 8th but not 7th and 9th grade students.</p> Conclusion <p>The <i>Movimente</i> Program was not effective in improving PA among the whole sample. Further studies are warranted to explore effective ways to improve PA among adolescents from Brazil.</p> Trial registration <p>The research project is registered with Clinical Trials (NCT02944318 Date of registration 25/10/2016).</p>