This study investigated the adhesion of resin composite to mineral trioxide aggregate based cements after different chemical and physico-chemical surface conditioning methods. Mineral trioxide aggregate based cements (Biodentine, ProRoot MTA, Imicryl MTA) were embedded in Teflon disks (N = 180). After storing at 37 degrees C at 100% humidity for 72 h, substrate surfaces were polished using silicon carbide papers. Specimens were allocated to 3 groups to be conditioned with one of the following (n = 15 per group): a) Adhesive resin (Clearfil SE Bond, CSE), b) Adhesive resin (Adper Single Bond 2, SB2), c) air-abrasion with 30 mu m alumina coated with silica + silane + adhesive resin (ALB), d) no surface conditioning, control group (CON). Microhybrid resin composite (Filtek Z250) was applied on the conditioned substrate surfaces and photo-polymerized. After storage at 37 degrees C at 100% humidity for 24 h, adhesive interfaces were loaded under shear (1 mm/min) in a universal testing machine. After debonding failure types were analyzed. Data were analyzed using 2-way ANOVA and Tukey's test (alpha = 0.05). SBS results were significantly affected by surface conditioning (p < 0.05) and materials (p < 0.05). Interaction terms were significant (p < 0.05). Biodentine-ALB resulted in significantly higher SBS values (3.96 +/- 1.24) compared to those of other combinations, while ALB and SB2 resulted in no significant difference for ProRoot MTA and Imicryl MTA (p > .05). CSE (1.36 +/- 0.5- 1.98 +/- 0.76) did not significantly increase SBS for all MTA materials compared to the control group (0.8 +/- 0.52 - 2 +/- 0.91) (p > 9.05). While CON groups resulted in exclusively adhesive failures, ALB presented the highest incidence of mixed failures for all materials tested (60-100%).