Since the restoration or masticatory function is the most important aim of implants, it should be substituted for the role of natural teeth and deliver the stress to the bone under the continous load during function. In natural teeth, stress distribution can be obtained through enamel, dentin and cementum and the elasticity of the periodontal ligament play a role of buffering action. In contrast, implant prosthesis has a very unique characteristics that it delvers the load directly to bone through the implant and superstructure. This fact arise the needs to evaluate the stress distribution of the implant in the mechnical aspects, which has a similar role of natural teeth but different pathway of stress. With 3 kinds of implant in prevalent use, 2 types of experimental PEA implant models were made, axisymmetric and 2-dimensional type. In axisymmetric model, the stiffness of the part including the prosthesis and implant which extrude out of bony surface could be calculated with displacement of the superstructure un er 100N vertical load and then damping effects could be determined through this stiffness. In axisymmetric FEA model, load to the bone could be deduced by evaluation the stress distribution of the designed surface under the 100N vertical force and in 2-dimensional model, 100N eccentric vertical load and 20N horizontal loda. The result are as follows. 1. In every implant, stress to the bone tends to be concenturated on the cortical bone. 2. Though the stress of the cancellous bone is larger at the apex of implants, it is less compared with cortical bone. 3. Under 20N horizontal load, stress of the left and right sides of implant shows a symmetrical pattern. But under 100N eccentric vertical load, loaded side shows much larger stress value. 4. In the 1mm interface, stress distribution among implants tend to have a similar pattern. But under 20N horizontal load apposite side of being loaded shows less stress in IMZ. 5. In the case of screw type implant, stress tends to vary along with screw shape. 6. According to the result determined with microstrain, cancellous bone id generally under the condition of overload, while cortical bone is usually within the limitation of physiologic load. 7. In the Branemark implant, maximum stress to the cortical bone is larger than any other implant except for the condition of 20N horizontal force and 0.05mm interface. 8. Damping effects of implants is maximum in IMZ.