The current study tested the effects of an intervention based on the trans-contextual model (TCM) on secondary school PE students' sport injury prevention behaviour, and on theory-based motivational and social cognition mediators. Participants were PE students (N=1,168; Mage =13.322±1.045, range=12 to 16; female=51.721%) who participated in a 3-month cluster-randomised controlled trial. Schools were randomly assigned to a treatment group, in which PE teachers received training to be more supportive of psychological needs in teaching sport injury prevention, or a control group, in which PE teachers received no training. Participants completed survey measures of TCM variables and self-reported sport injury prevention behaviour at baseline and at 3-month post-intervention follow-up. The proposed TCM model exhibited adequate fit with the data, χ2 =143.080 (df=19), CFI=.956, TLI=.916, RMSEA=.078 [90% CI=.066 to .090], and SRMR=.058. We found positive, statistically significant direct intervention effects on changes in perceived psychological need support (β=.064, p=.020). We also found positive, significant direct (β=.086 to .599, p<.001) and indirect (β=.002 to .027, p=.020 to .032) intervention effects on changes in TCM variables and behaviours to prevent sport injuries. Our findings support the TCM as a useful framework for building an intervention for promoting sport injury prevention behaviours among secondary school students.