ObjectivesLate adolescence (15–19 years) is a period of heightened susceptibility to stress, but regular physical activity may attenuate reactivity to stressors. We aimed to explore the effects of physical activity intensity on older adolescents’ stress-reactivity and self-reported mental health. Design and MethodsThree-arm randomised controlled trial in New South Wales, Australia (April–June 2021). Thirty-seven older adolescents (16.1 ± 0.2 years, 59.5 % female) were randomised to: i) non-active control (CON), ii) light-intensity physical activity (LPA), or iii) moderate-to-vigorous-intensity physical activity (MVPA). Physical activity groups participated in 2 x 20-min sessions/week for 6 weeks. Salivary cortisol (sCort) reactivity to induced stress was assessed using the Trier Social Stress Test for Groups and quantified as area under the curve (sCortAUC; Primary outcome). Secondary outcomes included peak cortisol (sCortPeak), subjective-reactivity, perceived stress, and non-specific psychological distress. Group differences were assessed using multiple linear regression and quantified using Cohen’s d. ResultsNo statistically significant effects were observed for sCortAUC or sCortPeak and the pattern of effects for subjective-reactivity was inconsistent. Effects for self-reported mental health were also non-significant (p > .05 for all) but of meaningful magnitude, favouring LPA and MVPA over CON (d’s = −0.38 to −0.54). Delivery fidelity was high, satisfaction was moderate-to-high, and there was no evidence of harm. However, recruitment, retention for sCort measures, and adherence were lower than expected. ConclusionSuboptimal recruitment, retention, and adherence limited our ability to conclude on the effect of physical activity intensity on older adolescents’ sCort-reactivity to induced stress. We observed potentially meaningful effects on self-reported mental health for both physical activity conditions, which could be confirmed in a future powered trial.