Purpose: This study aimed to evaluate the role of the implant/abutment system on torque maintenance of titanium retention screws and the vertical misfit of screw-retained implant-supported crowns before and after mechanical cycling. Materials and Methods: Three groups were studied: morse taper implants with conical abutments (MTC group), external-hexagon implants with conical abutments (EHC group), and external-hexagon implants with UCLA abutments (EHU group). Metallic crowns casted in cobalt-chromium alloy were used (n = 10). Retention screws received insertion torque and, after 3 minutes, initial detorque was measured. Crowns were retightened and submitted to cyclic loading testing under oblique loading (30 degrees) of 130 ± 10 N at 2 Hz of frequency, totaling 1 × 106 cycles. After cycling, final detorque was measured. Vertical misfit was measured using a stereomicroscope. Data were analyzed by analysis of variance, Tukey test, and Pearson correlation test (P .05). Results: All detorque values were lower than the insertion torque both before and after mechanical cycling. No statistically significant difference was observed among groups before mechanical cycling. After mechanical cycling, a statistically significantly lower loss of detorque was verified in the MTC group in comparison to the EHC group. Significantly lower vertical misfit values were noted after mechanical cycling but there was no difference among groups. There was no significant correlation between detorque values and vertical misfit. Conclusions: All groups presented a significant decrease of torque before and after mechanical cycling. The morse taper connection promoted the highest torque maintenance. Mechanical cycling reduced the vertical misfit of all groups, although no significant correlation between vertical misfit and torque loss was found.