To evaluate the effects of different smear layers on the microtensile bond strength (µTBS) of a reference two-step self-etch adhesive and two universal adhesives. Mid-coronal dentin of 90 teeth was exposed and divided into three bur groups (coarse diamond, fine diamond, or tungsten carbide). Each bur-prepared group was further divided into three adhesive groups: Clearfil SE Bond (SE, Kuraray Noritake), Single Bond Universal (SB; 3M Oral Care), and G-Premio Bond (GP, GC). After adhesive application, 4-mm-thick resin composites were built up. Half of the teeth in each bur-adhesive group were used in immediate µTBS testing, and the others were tested after thermal aging (n = 5). Rectangular sticks were prepared using a low-speed diamond saw. For each tooth, 6 central sticks were used in the µTBS test. Statistical analysis was performed using three-way ANOVA and Bonferroni tests (α = 0.05). SE presented higher µTBS than universal adhesives and SB presented higher µTBS than GP regardless of dentin surface preparation and thermal aging (p ˂ 0.05). For SE and SB, the tungsten carbide bur demonstrated higher immediate and aged µTBS than did the extra-fine diamond bur (p ˂ 0.05). The immediate µTBS was similar for GP with all bur types (p ˃ 0.05); the tungsten carbide and extra-fine diamond burs presented higher µTBS than did the coarse-diamond bur after thermal aging (p ˂ 0.05). Dentin surface preparation and adhesive type had significant effects on µTBS. The smear layer created with an extra-fine diamond or tungsten carbide bur is favorable when mild and ultra-mild self-etch adhesives are used.