The evaluation of microleakage is a method conventionally used to evaluate the composite restoration in cavities prepared by laser or conventional drill. The samples were five sound molar human extracted teeth with cavity preparation Class V in each surface (buccal, lingual, mesial and distal) divided in: GI -- cavity preparation with Er:YAG laser, adhesive agent Single Bond (3M), composite Z100 (3M) color A2; GII -- cavity preparation laser, 35% phosphoric acid etching, adhesive, composite; GIII -- cavity preparation with a drill, adhesive, composite; GIV -- cavity preparation with a drill, acid etching, adhesive, composite. The samples were sealed with a coating of nail varnish except on the area of the restoration and immersed in 50% aqueous solution of silver nitrate for 24 hours while kept in darkness. Samples were rinsed, soaked in photorevealing solution and exposed to fluorescent light for 6 hours. After, they were embedded in epoxy resin and sectioned longitudinally using the diamond saw microtone under running water. The specimens were observed under optical microscopy and the results were analyzed under Kruskall Wallis resulting in difference statistically between GI X GIV at the 5% level; and under Mann-Whitney resulting non-difference statistically between groups divided in cervical or occlusal walls.