Purpose: To evaluate the biomechanical behavior between two different micro conical abutments: two-piece and one-piece abutments. Materials and Methods: Four groups were divided according to the prosthetic connection (internal or external hexagon) and abutment type. The vertical misfit between the abutment and implant was evaluated under stereomicroscope. Next, the implants were divided into pairs and embedded in polyurethane to receive two-element prostheses. The initial removal torque was measured for abutments and for prosthesis fixation. The samples were mechanically cycled (200 N, 2 Hz, 2 × 10⁶ cycles) before measuring the torque loss in the same regions. For stress verification, a finite element analysis was used. Results: For vertical misfit, one-piece (4.70 ± 0.26 μm) two-piece (16.8 ± 0.32 μm). For preload in abutment screw, two-piece = one-piece. For prosthetic screw, a higher percentage of preload was found for two-piece, but no statistical difference was observed after fatigue cycling between the groups. For stress analysis, no difference (10%) was observed for abutments, retention screw, or for the bone. Conclusion: The one-piece abutment maintained even amounts of preload after cyclic loading and presented less vertical misfit in comparison with the two-piece abutment for both implant connections. The use of one-piece abutments did not increase the stress and strain concentration, suggesting acceptable biomechanical behavior. Keywords: abutments, finite element analysis, dental implants, implant-abutment interface, nonlinear analysis, torque loss