OBJECTIVE: To determine the microtensile bond strength and micromorphological structures on bonding of two adhesives (OptiBond FL and Clearfil SE Bond) to dentin surfaces ground with different preparation methods. METHODS: Extracted human molars were ground flat to expose mid-coronal occlusal dentin surface with one of six preparation methods--P120 grit SiC paper, P400 grit SiC paper, P1200 grit SiC paper, medium grit diamond bur, fine grit diamond bur, and carbide bur. Each of the adhesives was used to bond resin-based composite to the dentin surface. Dumbbell-shaped specimens were fabricated and microtensile bond strengths were determined. The subsequent debond pathway and micromorphological structures of representative dentin surfaces were examined under scanning electron microscopy. ANOVA and survival analyses were performed both assuming independence from and accommodating for within-tooth correlation between specimens. RESULTS: By ignoring the correlations between specimens, statistical analyses revealed no surface preparation effect on microtensile bond strength for each adhesive system. However, effects of surface preparation method on dentin adhesion of both adhesives were detected when accommodating for any within-tooth specimen correlations. Overall, carbide bur group showed the lowest bond strength for both OptiBond FL and Clearfil SE Bond. Dentin surfaces ground with diamond burs tended to present more compact smear layer than those ground with SiC papers and, subsequently, produced an effect on resin-dentin bond strengths. CONCLUSIONS: The dentin surface preparation method affects smear layer characteristics and dentin surface topography and, therefore affects resin-dentin bond strength. Smear layer denseness, more so than thickness, may compromise bonding efficacy of adhesives, especially of the self-etch systems.