Surface quality of cast metal is directly related to service performance under fatigue stress. Surface heterogeneities resulting from either finishing or polishing processes or by corrosive agents such as fluoridated solutions, can negatively affect fatigue life. Cast titanium frameworks are difficult to polish, and an accepted polishing protocol has not been established. This study evaluated and compared surface roughness of cast commercially pure titanium (CP Ti) and Ti-6Al-4V alloy submitted to conventional or electrolytic polishing, correlating the results with corrosion-fatigue strength testing performed in artificial fluoridated saliva. Specimens were also tested in air at room temperature to evaluate the effectiveness of the corrosion-fatigue test model. For each metal, 40 dumb-bell-shaped rods, 2.3 mm in diameter at the central segment, were cast. Conventional polishing was performed on 20 specimens of each metal following the manufacturer's instructions. A source of continuous electrical current was used for electrolytic polishing of the other 20 specimens of each metal, which were immersed in an electrolytic solution containing 5% fluoridric acid, 35% nitric acid, and 60% distilled water. Surface roughness, Ra (microm), was measured with a profilometer, and fatigue tests were carried out with a universal testing machine using a load 30% lower than the 0.2% offset yield strength. After failure, the fractured surfaces were examined using scanning electron microscopy. Surface roughness means were analyzed with a 2-way analysis of variance and the Tukey multiple comparisons test (alpha=.05). Electrolytic polishing (0.24 +/- 0.05 microm) provided significantly (P <.05) lower surface roughness values than conventional polishing (0.32 +/- 0.06 microm). Regardless of the polishing protocol, surface roughness of Ti-6Al-4V alloy (0.25 +/- 0.06 microm) was significantly lower (P <.05) than that of CP Ti (0.31 +/- 0.05 microm), and the fluoridated environment did not influence fatigue performance. There was no correlation between fatigue performance and surface roughness. Surface roughness of Ti-6Al-4V was significantly lower than that of CP Ti. For cast titanium frameworks, the electrolytic polishing regimen was found to be more effective than the manufacturer's polishing instructions with abrasives and rotary instruments. After polishing, differences in surface roughness values did not affect corrosion-fatigue performance.