The purpose of this study was to clarify the effects of applying different amounts of initial graft tension on the femorotibial positional relationship on the axial plane after anatomical ACL reconstruction. Eighty patients who underwent isolated ACL reconstructions using bone-patellar tendon-bone grafts were included in this study. In 40 of the 80 patients, the grafts were fixed at full knee extension with maximum manual force (high graft tension; Group H), whereas in the other 40 patients, the grafts were fixed at full knee extension with force of 80N (low graft tension; Group L). One week postoperatively, all patients underwent computed tomography (CT) on bilateral knee joints with knee extension. The femorotibial positional relationship in axial CT images were retrospectively evaluated. Side-to-side differences (the surgical side minus the unaffected side) were calculated in these variables. The side-to-side differences in anterior tibial translation distances were - 1.8 ± 2.1mm in Group H and - 1.9 ± 2.0mm in Group L, with no significant difference between the two groups. The side-to-side differences in tibial lateral shifts were - 0.2 ± 1.5mm in Group H and 0.0 ± 1.4mm in Group L, with no significant difference between the two groups. The side-to-side differences in tibial external rotation angles were 2.7 ± 4.5° in Group H and 0.3 ± 3.3° in Group L, with a significant difference between the two groups (P < 0.01). Applying high initial graft tension (maximum manual force) resulted in the external rotation of the tibia against the femur just after anatomical ACL reconstruction. In contrast, applying low initial graft tension (80N at full knee extension) did not change the femorotibial rotational relationship.