Abstract Background Recovery in mental health care comprises more than symptomatic improvement, but preliminary evidence suggests that only collaborative care may improve functioning of depressed older adults. This study therefore evaluates the effectiveness of behavioural activation (BA) on functional limitations in depressed older adults in primary care. Methods This study uses data from a multicentre cluster randomised controlled trial in which 59 primary care centres (PCCs) were randomised to BA and treatment as usual (TAU), and 161 consenting older (≥65 years) adults with clinically relevant symptoms of depression participated. Interventions were an eight-week individual BA programme by a mental health nurse (MHN) and unrestricted TAU. The outcome was self-reported functional limitations (WHODAS 2.0) at post-treatment (9 weeks) and at 12-month follow-up. Results At the end of treatment, the BA participants reported significantly fewer functional limitations than TAU participants (WHODAS 2.0 difference −3.62, p  = 0.01, between-group effect size = 0.39; 95% CI = 0.09–0.69). This medium effect size decreases during follow-up resulting in a small and non-significant effect at the 12-month follow-up (WHODAS 2.0 difference = −2.22, p  = 0.14, between-group effect size = 0.24; 95% CI = -0.08–0.56). MoCA score moderated these results, indicating that the between-group differences were merely driven by those with no cognitive impairment. Conclusions Compared to TAU, BA leads to a faster improvement of functional limitations in depressed older adults with no signs of cognitive decline. Replication of these findings in confirmatory research is needed.