Attention has been focused on the importance of anatomical tunnel placement in anterior cruciate ligament (ACL) reconstruction. The purpose of this study was to compare the effect of different tunnel positions for single-bundle (SB) ACL reconstruction on knee kinematics. Ten porcine knees were used for the following reconstruction techniques: three different anatomic SB [AM-AM (antero-medial), PL-PL (postero-lateral), and MID-MID] (n = 5 for each group), conventional SB (PL-high AM) (n = 5), and anatomic double-bundle (DB) (n = 5). Using a robotic/universal force-moment sensor testing system, an 89 N anterior load (simulated KT1000 test) at 30, 60, and 90 degrees of knee flexion and a combined internal rotation (4 N m) and valgus (7 N m) moment (simulated pivot-shift test) at 30 and 60 degrees were applied. Anterior tibial translation (ATT) (mm) and in situ forces (N) of reconstructed grafts were calculated. During simulated KT1000 test at 60 degrees of knee flexion, the PL-PL had significantly lower in situ force than the intact ACL (P < 0.01). In situ force of the MID-MID was higher than other SB reconstructions (at 30 degrees : 94.8 +/- 2.5 N; at 60 degrees : 85.2 +/- 5.3 N; and 90 degrees: 66.0 +/- 8.7 N). At 30 degrees of knee flexion, the PL-high AM had the lowest in situ values (67.1 +/- 19.3 N). At 60 and 90 degrees of knee flexion the PL-PL had the lowest in situ values (at 60 degrees : 60.8 +/- 19.9 N; 90 degrees : 38.4 +/- 19.2 N). The MID-MID and DB had no significant in situ force differences at 30 and 60 degrees of knee flexion. During simulated pivot-shift test at 60 degrees of knee flexion, the PL-PL and PL-high AM reconstructions had a significant lower in situ force than the intact ACL (P < 0.01). During simulated KT1000 test at 30, 60, and 90 degrees of knee flexion, the PL-PL and PL-high AM had significantly lower ATT than the intact ACL (P < 0.01). During simulated KT1000 test at 60 and 90 degrees, the MID-MID, AM-AM, and DB had significantly lower ATT than the ACL deficient knee (P < 0.01). During simulated KT1000 test at 90 degrees, every reconstructed knee had significantly higher ATT than the intact knee (P < 0.01). In conclusion, the MID-MID position provided the best stability among all anatomic SB reconstructions and more closely restored normal knee kinematics.