Objective. To determine the reproducibility and validity of the DynaPort KneeTest, a performance-based test that measures quality of movement of patients undergoing total knee replacement (TKR). Methods. A total of 92 patients with osteoarthritis (OA) of the knee performed the KneeTest twice on the same day; 94 healthy controls performed the KneeTest once. During the test, 29 activities were performed with accelerometers attached to the body. Relevant functional parameters were extracted from the accelerometers. A selection of parameters was used to calculate activity scores, based on the ability of parameters to discriminate between patients and controls (regression analyses). Based on internal consistency analyses (Cronbach's α), redundant activities were removed. Four cluster scores and a total KneeScore were calculated from the remaining activity scores. Reproducibility and validity of the cluster scores and the total KneeScore 2 were assessed. Results. Based on internal consistency analyses, the test was reduced to 23 activities. Inter- and intraobserver reliability using intraclass correlation coefficients were 0.90 (0.83-0.94) and 0.95 (0.83-0.98), respectively. Limits of inter- and intraobserver agreement were -8.3 to 11.3 and -4.2 to 9.0. Construct validity was confirmed by expected correlations with the Western Ontario and McMaster University Osteoarthritis Index physical functioning (0.55), Medical Outcomes Study Short Form-36 Health Survey physical functioning (0.62), and Knee Society Score function (0.64). Conclusion. The KneeTest is a useful performance-based measure for research in patients with knee OA undergoing TKR, with good reliability and validity. Further research is required to improve its usefulness for clinical practice.