Objective: Exercise is an efficacious stand-alone therapy for mild-to-moderate depression, but little is known about the influence of physical activity levels on responses to depression treatment. This study aimed to prospectively assess the association between self-reported habitual physical activity levels and depression severity following a 12-week intervention. Method: 629 adults (75% women; aged 18-71 years) with mild-to-moderate depression were recruited from primary care centres across Sweden and treated for 12 weeks. The interventions included internet-based cognitive behavioural therapy (ICBT) and 'usual care' (CBT or supportive counselling). One third of all participants were taking anti-depressant medication. The primary outcome was the change in depression severity assessed using the Montgomery-Åsberg Depression Rating Scale (MADRS). Habitual physical activity levels were self-rated and based on the estimated frequency, duration and intensity of total physical activity, including planned exercise, 'during a typical week'. Prospective associations were explored using linear regression models (percentage change) with 95% confidence intervals (CI's). Results: Following adjustment for relevant covariates, high levels of habitual physical activity were associated with larger relative reductions in depression severity compared to low physical activity (β = - 9.19, 95% CI = - 18.46, - 0.09, p = 0.052) and moderate physical activity (β = - 10.81, 95% CI = - 21.09, - 0.53, p <0.05), respectively. Conclusion: Adults who routinely engage in high levels of physical activity respond more favourably to CBT-focused depression treatments than adults who engage in low-to-moderate levels of activity. The optimal level of physical activity associated with reductions in depression severity corresponds to consensus recommendations for maximizing general health. One limitation is the use of self-reported physical activity data.