The objective of this work was to present a method for controlling the mechanical state of the knee during fixation in an ACL reconstruction; that is adjusting the unloaded length of the graft at the time of fixation so that the force-carrying characteristics of the graft match that of the intact ACL, when the normal and reconstructed knees are subjected to the same external load. To demonstrate the new method, an experimental knee testing system was used to measure ligament and graft forces with buckle transducers, as external loads were applied to five fresh knee specimens with a pneumatic load apparatus. ACL and collateral ligament force data were collected as anteriorly directed loads of 90 N (20 pounds) were applied at 0°, 30°, 60°, and 90° of flexion in the normal intact knee and the same knee following ACL excision. A composite graft was prepared, consisting of distally based semitendinosus and gracilis autogenous tissue augmented with the Kennedy Ligament Augmentation Device. Femoral fixation was obtained using a fixation plug that was inserted through the lateral femoral condyle. With a 90 N anterior tibial load applied at 30° of flexion and a single buckle transducer measuring the total composite graft force, the graft length was varied using the fixation plug until the force matched that seen by the intact ACL under the same loading condition. The graft was then clamped in the fixation plug and the anterior load sequence at the four flexion angles was applied to this reconstructed knee state. Ligament and graft forces were measured and compared to the normal intact knee and cut ACL force data. The proposed force-setting method of ACL reconstruction resulted in graft forces that were closer to normal than when using current reconstruction methods based on isometry and pretension, and allowed better reproducibility of total graft forces among series of knees.