The purpose of this study was to determine whether the transplantation of human cells from a non-reattached injured anterior cruciate ligament (ACL) remnant could enhance tendon-bone healing. Human ACL remnant tissue was classified into two groups based on the morphologic pattern as per Crain's classification:​ (1) non-reattachment group (Crain Ⅳ) and (2) reattachment group (Crain Ⅰ–Ⅲ). Seventy-five 10-week-old immunodeficient rats underwent ACL reconstruction followed by intracapsular administration of one of the following:​ (1) ACL-derived cells from the non-reattached remnant (non-reattachment group) (n = 5), (2) ACL-derived cells from the reattached tissue (reattachment group) (n = 5), or (3) phosphate-buffered saline (PBS) only (PBS group) (n = 5). Histological (Weeks 2, 4, and 8), immunohistochemical (Week 2), radiographic (Weeks 0, 2, 4, and 8), and biomechanical (Week 8) assessments were performed. Histological evaluation showed high and early healing, induction of endochondral ossification-like integration, and mature bone ingrowth at Week 4 in the non-reattachment group. Microcomputed tomography at Week 4 showed that the tibial bone tunnels in the non-reattachment group were significantly reduced compared to those in the reattachment and PBS groups. Moreover, biomechanical testing showed that ultimate load-to-failure in the non-reattachment group tended to be larger than that in the reattachment group, though not statistically significant. The enhanced healing potential in the non-reattachment group was explained by the increase in intrinsic angiogenesis/osteogenesis. In the subacute phase, the ACL-derived cells with the non-reattached morphologic pattern showed greater and earlier tendon bone healing compared with the cells obtained from the reattached morphologic pattern.