The goal of this study was to follow a cohort of patients undergoing total knee arthroplasty over time to: (1) identify and describe the various pain trajectories beginning preoperatively and for up to 12 months after surgery, (2) identify baseline predictors of trajectory group membership, and (3) identify trajectory groups associated with poor psychosocial outcomes 12 months after surgery. One hundred seventy-three participants (female = 85 [49%]; mean age [years] = 62.9, SD = 6.8) completed pain and psychological questionnaires and functional performance tests preoperatively and 4 days, 6 weeks, and 3 and 12 months after total knee arthroplasty. Using growth mixture modeling, results showed that a 4-group model, with a quadratic slope and baseline pain data predicting trajectory group membership, best fit the data (Akaike information criterion = 2772.27). The first 3 pain trajectories represent various rates of recovery ending with relatively low levels of pain 12 months after surgery. Group 4, the constant high pain group, comprises patients who have a neutral or positive pain slope and do not show improvement in their pain experience over the first year after surgery. This model suggests that preoperative pain levels are predictive of pain trajectory group membership and moderate preoperative pain, as opposed to low or high pain, is a risk factor for a neutral or positive pain trajectory postoperatively. Consistent with previous studies, these results show that postoperative pain is not a homogeneous condition and point to the importance of examining intraindividual pain fluctuations as they relate to pain interventions and prevention strategies.