To evaluate the tensile bond strength of two self-etching adhesive systems and one one-bottle system applied on dentin surfaces after different smear layer treatments. The crowns of 36 extracted third molars were transversally sectioned with a diamond saw to expose dentin. Flat dentin surfaces were abraded wet with 600-grit SiC paper and randomly assigned to 12 groups, according to three adhesive systems - Prime & Bond NT (PB), Clearfil Mega Bond (CMB) and Etch & Prime 3.0 (EP) - and four dentin surface treatments: direct application over smear layer (no treatment), sandblasting with 50 µm Al2O3 for 10 s, etching with 36% phosphoric acid for 15 s, or conditioning with 0.5 M EDTA for 2 min. Composite "crowns" (6.0 mm high) were incrementally built on the bonded surfaces and teeth were stored in distilled water at 37°C. Teeth were vertically, serially sectioned in both x and y directions to obtain bonded specimens of approximately 0.8 mm2 cross-sectional area. Each specimen was tested in tension in a universal testing machine at 0.5 mm/min. Overall microtensile bond strength (MPa) was 40.43 ± 7.63 for PB, 27.03 ± 8.40 for CMB, and 14.73 ± 3.91 for EP. Mean bond strength for PB was significantly higher than self-etching systems. No significant differences were observed among smear layer treatments within the same dentin adhesive. Bond strength values were significantly different according to the adhesive system used. Different smear layer treatments prior to bonding did not affect tensile bond strength within the same bonding agent.