ObjectiveTo explore the mediators of effects of two 6‐month telehealth‐delivered exercise programs, including exercise with and without weight‐loss diet, on pain and function improvements in knee osteoarthritis (OA).MethodsSecondary analysis of 345 participants from a 3‐arm randomized controlled trial of exercise (Exercise program) and exercise plus diet (Diet + Exercise program) versus information (Control program) was conducted. Outcomes were changes in pain (11‐point numeric rating scale) and function (Western Ontario and McMaster Universities Osteoarthritis Index [score range 0–68]) at 12 months. Potential mediators were change at 6 months in attitudes toward self‐management, fear of movement, arthritis self‐efficacy, weight, physical activity, and willingness for knee surgery. For the Diet + Exercise program versus the Exercise program, only change in weight was evaluated.ResultsPossible mediators of the Exercise program versus the Control program included reduced fear of movement (accounting for –1.11 units [95% confidence interval (95% CI) –2.15, –0.07] improvement in function) and increased arthritis self‐efficacy (–0.40 units [95% CI –0.75, –0.06] reduction in pain, –1.66 units [95% CI –3.04, –0.28] improvement in function). The Diet + Exercise program versus the Control program mediators included reduced fear of movement (–1.13 units [95% CI –2.17, –0.08] improvement in function), increased arthritis self‐efficacy (–0.77 units [95% CI –1.26, –0.28] reduction in pain, –5.15 units [95% CI –7.34, –2.96] improvement in function), and weight loss (–1.20 units [95% CI –1.73, –0.68] reduction in pain, –5.79 units [95% CI –7.96, –3.63] improvement in function). Weight loss mediated the Diet + Exercise program versus the Exercise program (–0.89 units [95% CI –1.31, –0.47] reduction in pain, –4.02 units [95% CI –5.77, –2.26] improvement in function).ConclusionIncreased arthritis self‐efficacy, reduced fear of movement, and weight loss may partially mediate telehealth‐delivered exercise program effects, with and without diet, on pain and/or function in knee OA. Weight loss may partially mediate the effect of diet and exercise compared to exercise alone.