To describe the natural history of pain after total knee arthroplasty and to identify factors predicting excessive postoperative pain, we used a prospective, observational study assessing clinical and radiographic variables preoperatively and at 1, 3, 6, and 12 months after knee replacement. Data sources included the visual analog pain scale and other measures of patient health, psychologic state, and component reliability. Regression analyses were conducted to identify specific factors predictive of postoperative pain, controlling for inequality of variables, and confirmed using regression diagnostics. For 116 patients (149 knees; mean age, 66 years; 55.2% women), significant pain was reported by 72.3%, 44.4%, 22.6%, 18.4%, and 13.1%, respectively. No intergroup differences existed for anesthesia, weight, age, or gender. Patients with greater preoperative pain had more postoperative pain, used more home therapy, and postoperative manipulations. Preoperative depression and anxiety were associated with heightened pain at 1 year. Pain after knee replacement resolves quickly, declining to approximately (1/2) by 3 months. However, one in eight patients report moderate to severe pain 1 year after surgery despite an absence of clinical or radiographic abnormalities. Development of office-based preoperative screening tools and interventions for these patients may reduce postoperative costs and improve patient-perceived outcomes.