In the experimental model, an implant cavity 3.0 mm in diameter was formed prior to embedding an implant 3.75 mm in diameter. In theory, the threads were completely mechanically fitted to the artificial mandibular bone. It does not osseointegrate, but does represent the circumstances of immediate loading in a state of full contact with the bone. The contact model reproduced the state of contact between the bone and implant in the experimental model; theoretically, displacements under loading conditions should show values equivalent to those in the experimental model. Nonetheless, the displacement under loading conditions in the contact model had values 1/3 to 1/4 of those observed in the experimental model. There are three conceivable possibilities. One is that it is possible that consecutive loading of the superstructure on the buccal and lingual sides causes implant loosening. In such a case, displacement under loading conditions would be larger than in the experimental model. Sato et al. [26] reported that the yield tensile load of a screw was 656 N. The fatigue limit causing screw loosening or fracture is half of the yield tensile load, or 328 N. Using a geometric analysis, the largest tensile force in the gold screw after a buccal loading of 100 N was 73 N. In this reported case, if more than 450 N was applied to the loading point, the screw will loosen or fracture. Therefore, 100 N of consecutive superstructure loading on the buccal and lingual sides was not a cause of implant loosening. Furthermore, it is not a reason for displacement under loading conditions in the experimental model to be greater. The second reason is that the experimental model was simply placed on the worktable. Therefore, there is the possibility that minute movements occurred during the vertical loading process. Additionally, measured implant displacement under loading may have overestimated the actual displacement. In order to reproduce complete constraint conditions in FEA models under loading conditions, the bottom surface of the artificial mandibular bone should be adhered completely to the worktable with adhesive; this should stabilize the experimental model and minimize minute movements during the loading process. The third reason is that it is thought that the Young’s modulus assigned to the FEA models was different from the actual Young’s modulus of artificial mandibular bone. In an FEA, the physical properties assigned to elements reportedly have a major impact on the analysis results [27,28]. Nomura et al. [29] reported that displacement under loading conditions increases when the Young’s moduli of the cortical and cancellous bone are reduced. We used the manufacturer’s publicly disclosed values for the Young’s modulus of the artificial mandibular bone, but it is not clear how these values were measured. In particular, with respect to artificial cancellous bone, the interior has become a foam state and it is thought that the Young’s modulus is smaller than the publicly disclosed value. In such a case, displacement under loading conditions would be greater than measured in this study’s FEA models and would be nearer to the displacement under loading conditions in the experimental model. An accurate method for measuring the Young’s modulus also requires future study.