A radial head fracture is a common fracture of the elbow joint most frequently caused by a fall on an outstretched arm. When the radial head is comminuted and reconstruction is not viable, radial head replacement is often the preferred option. There are basically two choices for the radial head implants: monoblock and bipolar. However, little research examining the performance of available radial head implants has been conducted. In this study, a finite element model for the intact elbow joint extracted from CT scans of a cadaveric elbow was created. The constructed FE model was validated with cadaveric experiments using Fuji pressure sensitive film. Then, two different implant models were generated by replacing the radial head with monoblock and bipolar implants. Contact stress and area for each model were calculated and compared under various loading conditions in the contact areas between the radial head and the capitellum of the humerus, and between the radial head and the radial notch of the ulna. A significant difference was found in both maximum contact stress and contact area between the native elbow and two types of implants. Contact stress for monoblock was higher than that for bipolar and vice versa for contact area. FE predictions showed the same pattern as measurements from Fuji analysis did although they overestimated the maximum contact stress and underestimated the contact area compared to the Fuji film experimental results.