Osteoarthritis (OA) is a common joint disorder for which there is no cure. Current treatments are suboptimal. Exercise is a core treatment for knee OA, with muscle strengthening exercise commonly recommended. Yoga is a mind-body exercise intervention that can improve flexibility, muscle strength, balance, and fitness and potentially reduce symptoms of OA. However, there is a scarcity of robust, high-quality conclusive evidence on the efficacy of yoga in knee OA. We are currently conducting the first randomised comparative effectiveness and cost-effectiveness trial of a yoga program compared with a strengthening exercise program in patients with symptomatic knee OA. This study protocol describes the design and conduct of this trial. The YOGA study is a phase III, single-centre, parallel, superiority, randomised, active-controlled trial which will be conducted in Hobart, Australia. One hundred and twenty-six participants (63 in each arm) aged over 40 years with symptomatic knee OA will be recruited from the community and randomly allocated to receive either a 24-week yoga program (3×/week) or a strengthening exercise program (3×/week). The primary outcome will be change in knee pain over 12 weeks, assessed using a 100 mm visual analogue scale (VAS). The secondary outcomes include change in knee pain, patient global assessment, physical function, quality of life, gait speed, biomarkers, and others over 12 and 24 weeks. We will also assess whether the presence of neuropathic pain moderates the effects of yoga compared to strengthening exercise. Additional data, such as cost and resource utilization, will be collected for the cost-effectiveness analysis. The primary analysis will be conducted using an intention-to-treat approach. Adverse events will be monitored throughout the study. Once completed, this trial will contribute to the knowledge of whether yoga can be used as a simple, effective, low-cost option for the management of knee OA, thus saving economic costs in the healthcare system.