Philosophiae Doctor - PhD / Background: The Human Immunodeficiency virus (HIV) has become a global public health challenge amid the growing concern of adolescent risky sexual behaviour, influenced by biological and psycho-social factors. There is an increasing demand for adolescent sexual risk reduction interventions, especially in sub-Saharan Africa which remains the hub of HIV epidemic worldwide. Sport-based HIV prevention programme has been identified as one of the interventions, in addition to other approaches to provide young people with appropriate HIV knowledge and skills to either delay or reduce risk-taking behaviour. Despite the potential of sport-based programme for sexual risk reduction, there is limited information on how to adapt it to meet adolescent needs in terms of design, contents, and delivery. This information is necessary to provide sufficiently strong evidence to support widespread implementation of sport-based programme, especially in rural African schools. Therefore, the study aimed to measure the impact of a sport-based HIV prevention intervention in the reduction of HIV related sexual risk behaviour among rural high school learners aged 13-18. Method: The overall study design was a concurrent mixed method, utilizing both the quantitative and qualitative approach. The population for the quantitative and the qualitative studies was made up of two high schools in a predominantly “Coloured” community in Ceres. Two classes each were randomly selected from grades 8-10 in both intervention and control school. The intervention was developed through a process of focus group discussions with the Grassroot Soccer (GRS) staff and it was guided by the Social Cognitive Theory. The intervention was delivered to grade 8-10 learners by GRS peer facilitators using the GRS generation skillz curriculum that consists of 12-week sessions in the intervention school. Quantitative data were analysed with multivariate statistical techniques and qualitative data with thematic analysis approach. Results: The data that assessed the behavioural and protective factors to understand why learners in rural schools engage in risky sexual behaviour showed that about 27.2% of the learners reported being sexually active. Of the sexually active learners, 48.7% reported engaging in sex by 14 years or younger and nearly 42.2% reported multiple sexual partners with significant higher proportion of boys than girls. Nearly 55.2% of the sexually active learners reported irregular condom use and 46.3% did not use a condom at the last sexual encounter. The majority of the learners (87%) did not know their HIV status. Being male (OR = 6.60;95% C I = 1.62 – 26.84) and peer influence (OR = 3.01; 95% CI = 1.97-4.60) were the strongest predictors of reporting sexual intercourse and early sexual activities before the age 15 respectively. Though the knowledge of HIV was low, those with greater knowledge of HIV were more likely to use a condom at last the sexual encounter (OR = 1.22; 95% C I =1.03-1.44). The learners who participated in sport-based intervention were 1.43times likely to report higher self-efficacy to refuse sex compared to the control group (OR = 1.43; 95% C.I =1.07-1.92).The process evaluation indicated that the sport-based intervention was well received among the learners as it gave them free space to freely express themselves. Conclusion: The findings have shown that sport-based intervention can be successfully implemented in school and is a promising approach to reduce risks associated with risky sexual behaviour in learners. However, the quality of the programme delivery was hampered by irregular session‟s schedule and language of instruction. The process suggests for a longer exposure period and because of social-cultural diversity, learners must be provided with the ownership of the programme in schools.