The purpose of this study was to evaluate Patient-Reported Outcomes Measurement Information System physical function (PROMIS PF) 2 years following knee surgery, and identify preoperative factors associated with postoperative PROMIS PF. Three hundred and sixty-five patients, age 17 years and older, undergoing knee surgery at one institution were studied. Patients completed multiple questionnaires prior to surgery and again 2 years postoperatively including PROMIS PF, International Knee Documentation Committee (IKDC), joint and body numeric pain scales (NPS), Tegner’s activity scale (TAS), and Marx’s activity rating scale (MARS). Mean PROMIS PF improved from 41.4 to 50.9 at 2 years postoperatively ( < 0.001) and was strongly correlated with 2-year IKDC scores. Older age, female gender, non-Hispanic ethnicity, unemployment, lower income, government insurance, smoking, preoperative opioid use, having a legal claim, comorbidities, previous surgeries, higher body mass index (BMI), and knee arthroplasty were associated with worse 2-year PROMIS PF. Multivariable analysis confirmed that lower BMI, less NPS body pain, and higher MARS were independent predictors of greater 2-year PROMIS PF and better improvement in PROMIS PF. In this large, broad cohort of knee surgery patients, multiple preoperative factors were associated with PROMIS PF 2 years postoperatively. PROMIS PF scores improved significantly, but worse 2 year PROMIS PF scores and less improvement from baseline were independently predicted by higher BMI, greater NPS body pain, and lower MARS activity level. PROMIS PF can be implemented as an efficient means to assess outcomes after knee surgery.Thieme Medical Publishers 333 Seventh Avenue, New York, NY 10001, USA.