This study evaluated the effect of two desensitizing procedures on the marginal microleakage of composite restorations bonded by a two-step etch-and-rinse adhesive. Class V cavities were prepared on the buccal surfaces of 42 extracted premolars at the cementoenamel junction and randomly divided into three groups of 14 each. In group 1 (control), an adhesive (Excite) was applied according to manufacturer's instructions. In group 2, after the application of oxalate desensitizer (BisBlock) to the cavities, the specimens were etched, and Excite was applied and light cured. In group 3, the cavities were treated with CO(2) laser irradiation (1 W power output in a continuous wave mode, 10 s), and then, the specimens were etched and bonded by Excite. All the cavities were restored with a resin composite. After water storage and thermocycling, the specimens were placed in 1 % basic fuchsin blue dye solution. The dye penetration was evaluated using a stereomicroscope. The data were analyzed by using Kruskal-Wallis and Mann-Whitney statistical tests (α = 0.05). There was a significant difference among three groups at the gingival and occlusal margins (P < 0.05). The specimens treated with BisBlock exhibited higher microleakage at both margins than those of the control and laser-treated groups (P < 0.001). There was no significant difference between the control group and group treated with CO(2) laser at both margins (P > 0.05). CO(2) laser did not show a detrimental effect on the marginal sealing ability of the adhesive system used in resin composite restorations. However, oxalate desensitizer adversely influenced the sealing ability of the restorations.