Background: Control cognitions have been directly related to positive engagement with rehabilitation regimes. The impact of such cognitions on recovery following surgery is not well understood. Purpose: To assess whether perceived control cognitions predict function 9–12 months following total hip replacement (THR). Methods: Prospective cohort study performed as part of a randomised controlled trial. Behavioural cognitions (BC) (recovery locus of control (RLOC); perceived external behavioural control (PEBC))) and subjective functional outcome measures (Oxford hip score (OHS) and a reduced version of the Western Ontario and McMasters University Osteoarthritis Function scale (rWOMAC PF)) were administered pre-operatively and up to 12 months post-operatively to 50 patients randomised to home-based progressive resistance training (N = 26) or standard rehabilitation (N = 24), post-THR. Regression analysis investigated variance in functional scores. Results: Group randomisation had no effect on BC. RLOC and OHS (6 months) correlated significantly with 12-month OHS, with 6-month OHS predicting 62.3% of the variance in 12-month OHS. 12-month rWOMAC PF was determined by each of its three previous assessments (pre-operative 8.8%, 6 weeks 17.8% and 6 months 67.3%). Variance in functional gain at 12 months (OHS and rWOMAC PF) was explained by pre-operative OHS and rWOMAC PF (63.7% and 63.8%, respectively). Conclusions: BC had no impact on functional outcome in this population. Subjectively assessed function at 12 months, as well as the levels of functional gain over time, was best explained by the patients’ earlier functional status.