Purpose: This in vitro study was designed to compare the effects of oxygen plasma and sandblasting treatment of the surface of titanium implant abutments on retention of implant-supported fixed prostheses using eugenol-free provisional cement and surface topography. Cement failure mode and elemental composition were surveyed. Materials and Methods: Thirty implant abutments were used. They were divided into three groups (n = 10 implants) according to the surface treatment made (no treatment, sandblasting with 50 μm Al2O3, and oxygen plasma). Thirty metal copings were fabricated with an occlusal metal O-ring and cemented using eugenol-free cement. The tensile bond strength was recorded using a universal testing machine, cement failure was recorded, surface topography was determined using the optical method, and elemental composition was determined using energy-dispersive x-ray spectroscopy. Data from all the groups were compared with a one-way analysis of variance (ANOVA; P .05) and Tukey test. Results: The mean values of tensile bond strength of metal copings bonded to three implant-abutment surface groups with either control, sandblasting, or oxygen plasma were 138.8 ± 10.2, 261.3 ± 20.6, and 264.7 ± 12.8 N, respectively. The cements failed adhesively in the control and oxygen plasma groups and cohesively for sandblasting. The mean values of surface roughness were 1.1 ± 0.2, 2.3 ± 0.5, and 1.0 ± 0.1, respectively. The oxygen plasma-treated group evidenced a decreased level of C and increased levels of O2. Conclusion: The retention of cast crown copings to titanium implant abutments using eugenol-free provisional cement in the sandblasting and oxygen plasma groups was more improved than in the control group. Palabras clave: abutment surface modification, implant-abutment surface, plasma-surface modification, provisional cement retention