Through a range of hypothetical mechanisms physical activity may provide benefits in the prevention of substance use across the lifespan. The primary objective of this study was to evaluate the prospective association between physical activity and the initiation of current substance use among youth. The secondary objective examined if school connectedness mediated any prospective association. This longitudinal study included grade 9–12 students (n = 3244) with three years of linked data (2017–2020) from the COMPASS study. Multi-level logistic regressions were built to explore the association between changes in physical activity (meeting MVPA guidelines, outside of school sport participation, competitive school sport and non-competitive school sport) and the initiation of current substance use behaviours (cigarette, e-cigarettes, cannabis and binge drinking). Analyses examined if school connectedness mediated the association between physical activity and substance use. Results indicated that females who initiated 60 min of MVPA daily had greater than three times the odds of initiating current cigarette smoking (OR = 3.26 [1.45, 7.33]). Additionally, males who consistently participated in outside-of-school sport had greater than two times the odds of initiating current binge drinking (OR = 2.28 [1.43, 3.63]). Our mediation analysis demonstrated that sport participation may lower current e-cigarette and cannabis use through school connectedness. Overall, there was no evidence that physical activity confers broad, universal benefits in preventing substance use among Canadian youth. However, this is the first work to indicate that sport participation may provide a contextual experience that enhances school connectedness which in turn is associated with substance use prevention.