ObjectivesTo investigate whether participants with knee osteoarthritis classified as ‘more’ or ‘less’ physically active at baseline differ in characteristics and/or outcomes at baseline and at 3 and 12 months following the commencement of an education and exercise‐therapy program.MethodsProspective cohort study using the GLA:D® Australia registry. The University of California, Los Angeles Physical Activity Scale (UCLA) participant data dichotomised as ‘more’ (≥7) or ‘less’ active (≤6). Groups were compared using chi‐square (obesity [baseline only], comorbidity prevalence, medication consumption, fear of damage from physical activity); and linear mixed model regression (12‐item Injury Osteoarthritis Outcome Score [KOOS‐12], pain [visual analogue scale], health‐related quality of life [QoL] [EQ‐5D‐5L]) statistics, adjusted for age, sex and baseline physical activity at 3 and 12 months.ResultsWe included 1059 participants (70% female). At baseline, 267 (25%) were classified as ‘more’ active, increasing to 29% and 30% at 3 and 12 months, respectively. At baseline, compared to the ‘less’ active group, the ‘more’ active group had a lower proportion of participants who were obese (‘more’ = 21% vs. ‘less’ = 44%), had comorbidities (58% vs. 74%) and consumed medications (71% vs. 85%); lower pain intensity (37 vs. 47); and higher KOOS‐12 (59 vs. 50), and health‐related QoL (0.738 vs. 0.665) scores. When accounting for age, sex and baseline physical activity, improvements seen in knee‐related burden and health‐related QoL were not different between groups at 3 or 12 months. Compared to the ‘less’ active group, the proportion of participants not consuming medication remained higher in the ‘more’ active group at 3 (‘more’ 45% vs. ‘less’ 28%) and 12 months (43% vs. 32%).Conclusion‘More’ active people with knee osteoarthritis were less likely to be obese, had fewer comorbidities, lower medication consumption, knee‐related burden and pain intensity, and higher health‐related QoL than ‘less’ active participants at all timepoints.