Background: Physical activity levels are low in people with chronic obstructive pulmonary disease and there is limited knowledge about how pulmonary rehabilitation transforms movement behaviours. This study analysed data from a pulmonary rehabilitation trial and identified determinants of movement behaviours. Methods: Objectively-assessed time in daily movement behaviours (sleep, sedentary, light intensity [LIPA], moderate to vigorous intensity [MVPA]) from a randomised controlled trial (n=73 participants) comparing home- and centre-based pulmonary rehabilitation was analysed using conventional and compositional analytical approaches. Regression analysis was used to assess relationships between movement behaviours, participant features and response to the interventions. Results: Compositional analysis revealed no significant differences in movement profiles between the home- and centre-based groups. At end rehabilitation, conventional analyses identified positive relationships between exercise capacity (6-minute walk distance, 6MWD), LIPA and MVPA time. Compositional analyses identified positive relationships between 6MWD and MVPA time, accompanied by negative relationships with sleep and sedentary time (relative to other time components) and novel relationships between body mass index and LIPA/sedentary time. Conclusion: Compositional analyses following pulmonary rehabilitation identified unique associations between movement behaviours that were not evident in conventional analyses.