The purposes of this study were to determine the relation between the graft tension and the knee flexion angle in the anteromedial (AM) and posterolateral (PL) bundles of a clinically simulated anatomic double-bundle anterior cruciate ligament (ACL) reconstruction procedure and to clarify the effect of initial tension on the tension-versus-flexion curve of each graft, as well as the effect of internal rotation of the tibia on the tension of the 2 grafts. During ACL reconstruction in 30 patients, 2 suture anchors with a No. 1 polyester suture were firmly screwed into the center of the anatomic attachment of the AM and PL bundles on the femur, respectively, and each graft tension was measured with a strain gauge-type tensiometer attached at the end of the suture under 2 conditions of initial tension. The averaged tension-versus-flexion curves were significantly different between the AM and PL grafts under each initial tension condition (P < .0001). The initial tension applied at 30 degrees of knee flexion significantly affected the absolute values of each graft tension at each knee flexion angle (P < .0001) but did not significantly affect the tension-versus-flexion curve pattern of each graft. The maximal internal rotation of the tibia significantly increased the tension on both the AM and PL grafts at knee flexion angles of less than 60 degrees under each initial tension condition (P < .0001). The tension-versus-flexion curves of the 2 sutures, which mimicked the AM and PL grafts reconstructed clinically with the anatomic double-bundle ACL reconstruction procedure, were significantly different in the tension values. Differences in initial tension applied to the 2 grafts significantly affected the absolute values of each graft tension at each knee flexion angle but did not significantly affect the tension-versus-flexion curve pattern. The maximal internal rotation of the tibia significantly increased the tension on both the AM and PL suture grafts at knee flexion angles of less than 60 degrees . Level I, testing of previously developed diagnostic criteria in series of consecutive patients with universally applied gold standard.