Abstract Background Effective lifestyle interventions are needed to prevent noncommunicable diseases in low- and middle-income countries. We analyzed the effects of a school-based health promotion intervention on physical fitness after 28 months and explored if the effect varied with important school characteristics. We also assessed effects on screen time, physical activity and BMI. Methods and results We performed a cluster-randomized pair matched trial in schools in urban Ecuador. The intervention included an individual and environmental component tailored to the local context and resources. Primary outcomes were physical fitness (EUROFIT battery), screen time (questionnaires) and physical activity (accelerometers). Change in BMI was a secondary outcome. A total of 1440 grade 8 and 9 adolescents (intervention: n = 700, 48.6%) and 20 schools (intervention: n = 10, 50%) participated. Data of 1083 adolescents (intervention: n = 550, 50.8%) from 20 schools were analyzed. The intervention increased vertical jump (mean effect 2.5 cm; 95% CI 0.8-4.2; P = 0.01). Marginally insignificant, adolescents from the intervention group needed less time for speed shuttle run (intervention effect = −0.8 s, 95% CI −1.58-0.07; P = 0.05). The proportion of students achieving over 60 minutes of moderate-to-vigorous physical activity/day decreased over time with the change in proportion significantly less in the intervention schools (6 vs. 18 percentage points, P