Background: The aim of this study was to evaluate the effect of natural saliva contamination of dentin, on marginal and internal adaptation of resin-composite restorations performed with an etch-and-rinse adhesive and determine the efficiency of various decontamination techniques. Methods: A hundred and twenty human molars and premolars were randomly distributed into six groups (n=20). Enamel was removed and standardized dentin cavities (diameter: 3 mm, depth: 2 mm) were filled with a two-step etch-and-rinse adhesive and resin-composite under six different surface treatments: (1) adhesive application following manufacturers’ instructions (control); (2) dentin etching, 5-s saliva, 5-s air-dry, adhesive; (3) dentin etching, 5-s saliva, 10-s water-rinse, 5-s air-dry, adhesive; (4) adhesive application/light-cure, 5-s saliva, 5-s air-dry; (5) adhesive application/light-cure, 5-s saliva, 10-s water-rinse, 5-s air-dry; (6) adhesive application/light-cure, 5-s saliva, 10-s water-rinse, 5-s air-dry, adhesive reapplication. Wall adaptation was evaluated at the upper surface and at two consecutive depth levels of 0.5mm each, by measuring the length of the debonded margins relative to the cavity periphery (%DM) and the width of the maximum marginal gap (MG). Results: Statistically significant differences occurred among groups, presenting inferior adaptation after contamination took place. Group (1) at the upper surface presented the most satisfactory adaptation, whereas, groups (4) and (5) had the significantly poorest adaptation in all surfaces. Groups (2), (3), and (6) revealed no statistically significant differences in comparison to group (1). Conclusion: Saliva contamination resulted in deterioration of marginal and internal adaptation when it occurred after the adhesive application. Reapplication of the adhesive restored DM and MG values.