Objective: The purpose of this study was to examine social cognitive constructs as predictors of preoperative exercise (PE) in a sample of individuals waiting for total joint replacement (TJR) surgery using the framework of Bandura’s Social Cognitive Theory (SCT). Methods: Participants (N = 78) were individuals waiting for TJR at the two major urban centres on Vancouver Island, Canada who completed measures of the SCT (barrier self-efficacy, outcome expectancy, self regulation, task efficacy & sociocultural factors of pain, physical function and neighbourhood walking environment) framed for (PE). Results: Independent t-tests suggested no differences between type of surgery (hip versus knee), gender or age for PE (p<.05). Over half of the sample was considered inactive (55%) using a definition of physical activity as accumulating at least 30 minutes of exercise at a moderate or vigorous intensity at least 3 days per week in bouts of 10 minutes or more. Bivariate correlations relating to PE were significant (p<.05) between self regulation (SR) (.25), task efficacy for exercise (TEE) (.27) and pain (-.28). Hierarchical regression analysis revealed that SR (β=.17) and TEE (β=.20) explained 10% of the variance in PE behaviour, but were not significant predictors of PE independently. The addition of pain to the regression analysis added 4% of the explained variance, and remained the only significant predictor (p<.05) of Pe behaviour. Conclusions: SCT showed modest capability in predicting PE in this sample, suggesting further testing of theoretical models is warranted in this area. These findings highlight the influence of pain on exercise before TJR surgery, and support the importance of considering individual factors such as pain when designing targeted interventions to increase activity in this population.