Background: The impact of posterior cruciate ligament-substituting and posterior cruciate ligament-retaining devices on the range of motion of the knee following primary total knee arthroplasty is unclear. The primary objective of our study was to compare the range of motion of the knee over the first two postoperative years between subjects who had received the ligament-substituting design and those who had received the ligament-retaining design. Secondarily, pain, function, and health-related quality of life were compared between the two groups. Methods: We undertook a prospective randomized study of 100 patients with noninflammatory osteoarthritis who were undergoing primary total knee arthroplasty. Patients were randomized at the time of surgery to receive one of the two study prosthesis designs. They were evaluated preoperatively, at the time of hospital discharge, at three months postoperatively, and at up to two years postoperatively by a physical therapist who was blinded to the group allocation. Active knee flexion and extension, measured with a goniometer, were the primary outcome measures. Self-reported pain, function, and health-related quality of life were measured with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the RAND 36-Item Health Survey (RAND-36). Complications and revision rates were determined through hospital record review and at each patient evaluation. Results: Eighty percent (seventy-eight) of the ninety-eight available patients returned for the final assessment. At two years postoperatively, the mean difference between the groups with regard to knee flexion was 0.03° (95% confidence interval, −5.9° to 6.0°) and the mean difference in knee extension was 1.0° (95% confidence interval, −0.36° to 2.4°). Ninety-one (93%) of the ninety-eight available patients completed questionnaires at the time of the final assessment. The two treatment groups had similar pain, function, and quality-of-life scores and complication rates. One subject in the cruciate-retaining group required revision, secondary to a deep joint infection, and one subject in the cruciate-substituting group required manipulation to address poor knee flexion. Conclusions: Overall, the two treatment groups had a similar range of motion of the knee over the initial two-year postoperative time period. A satisfactory range of motion was achieved by three months postoperatively and was maintained at the final assessment. Level of Evidence: Therapeutic Level I. See Instructions to Authors for a complete description of levels of evidence.