New strategies to fulfill craniofacial bone defects have gained attention in recent years due to the morbidity of autologous bone graft harvesting. We aimed to evaluate the in vivo efficacy of bone tissue engineering strategy using mesenchymal stem cells associated with two matrices (bovine bone mineral and α-tricalcium phosphate), compared to an autologous bone transfer. A total of 28 adult, male, non-immunosuppressed Wistar rats underwent a critical-sized osseous defect of 5 mm diameter in the alveolar region. Animals were divided into five groups. Group 1 (n = 7) defects were repaired with autogenous bone grafts; Group 2 (n = 5) defects were repaired with bovine bone mineral free of cells; Group 3 (n = 5) defects were repaired with bovine bone mineral loaded with mesenchymal stem cells; Group 4 (n = 5) defects were repaired with α-tricalcium phosphate free of cells; and Group 5 (n = 6) defects were repaired with α-tricalcium phosphate loaded with mesenchymal stem cells. Groups 2-5 were compared to Group 1, the reference group. Healing response was evaluated by histomorphometry and computerized tomography. Histomorphometrically, Group 1 showed 60.27% ± 16.13% of bone in the defect. Groups 2 and 3 showed 23.02% ± 8.6% (p = 0.01) and 38.35% ± 19.59% (p = 0.06) of bone in the defect, respectively. Groups 4 and 5 showed 51.48% ± 11.7% (p = 0.30) and 61.80% ± 2.14% (p = 0.88) of bone in the defect, respectively. Animals whose bone defects were repaired with α-tricalcium phosphate and mesenchymal stem cells presented the highest bone volume filling the defects; both were not statistically different from autogenous bone.