This study evaluated the vertical misfit, passivity, and stress distribution after tightening the screws of different prosthesis. Two implants were used to simulate the rehabilitation of partially edentulous mandible space from the second premolar to the second molar. 40 three-element screw-retained fixed dental prosthesis with distal cantilever were fabricated and divided into four groups according to the method of production of framework (n = 10): G1 = conventional casting one-piece framework, G2 = conventional casting sectioned and laser welding, G3 = conventional casting sectioned and tungsten inert gas (TIG) welding and G4 = framework obtained by CAD/CAM (computer-aided design/computer-aided manufacturing) system. The vertical misfits (both screws tightened) and the passive fit (one screw tightened) were measured under a comparator optical microscope. The data was submitted to Shapiro-Wilk test to enable comparison with ANOVA followed by Tukey with Bonferroni adjust (α = .05). The qualitative analysis of the stress distribution was performed by the photoelastic method. The vertical misfit (both screws tightened) of the G2 (24 μm) and G3 (27 μm) were significantly higher than G4 (10 μm) (p = 0,006). The passive fit (for the non-tightened) of the G1(64 μm) and G3 (61 μm) were significantly higher than the G4 (32 μm) (p=0,009). G1 showed high stress between the implants in the photoelastic analysis and G4 presented lower stress. In conclusion, CAD/CAM method results in less vertical misfit, more passivity, and consequently better stress distribution to the bone.