OBJECTIVES This study evaluated the marginal adaptation of composite indirect restorations bonded with dual curing resin cement after different strategies to seal dentin. Different bonding techniques associated or not with a low-viscosity composite resin (LVCR) were utilized. In addition, the bond strength between composite resin and pre-sealed dentin was evaluated in the buccal and pulpal walls of class I cavities, prepared for indirect restorations. METHODS Thirty-three freshly extracted human molars were used for this study, divided into three groups (n=11) representing different techniques to seal dentin-(Group 1) Conventional technique: the adhesive system was applied and polymerized just before the cementation of the indirect restoration; (Group 2) Dual bonding technique: a first layer of the adhesive system was applied and polymerized just after preparation, and a second layer just before the final cementation; (Group 3) Resin coating technique: a LVCR was applied and polymerized after the first layer of the adhesive system, and before the impression. A further application of the adhesive system was performed before the placement of the restoration. The restorations were polished and a solution of acid red propylene-glycol was dropped on each specimen's occlusal surface for 10 s. The dye penetrations were captured under stereoscopic lens and the images were transferred to a computer with a measurement program, in order to determine the extension of the dye penetration. The microtensile bond strength test (muTBS) was applied on pulpal (P) and buccal (B) walls of the restorations for Groups 1-3. The subgroups for muTBS were: Group 1P (n=13); Group 1B (n=7); Group 2P (n=6); Group 2B (n=14); Group 3P (n=14); Group 3B (n=15). All specimens were sectioned to obtain an area of 0.8 mm2. The specimens were mounted on a microtensile device and fractured using a universal testing machine at a cross-head speed of 1mm/min. Failure modes were analyzed by SEM. One-way ANOVA and multiple-comparison Tukey's test were used for statistical analysis of the marginal adaptation scores and muTBS test. Non-parametrical Kruskal-Wallis test was used for failure mode analysis. RESULTS Group 3 showed a significantly higher mean value of marginal dye penetration (45.59) when compared to Groups 1 (8.44) and 2 (18.92). For pulpal walls, Group 1P showed significantly higher mean muTBS (25.93+/-2.27) when compared to Groups 2P (14.71+/-1.78) and 3P (16.07+/-2.81). There was no statistical difference between Groups 2P and 3P. For buccal walls, Group 2B presented significantly higher mean muTBS (23.29+/-1.42), and Group 1B the lowest mean values (11.37+/-1.14). The failure mode analysis presented a considerable variation, according to the treatment and to the wall. SIGNIFICANCE The results of this study indicated that a previous sealing of dentin using the adhesive system, followed by a second application just before the cementation, is an effective alternative technique, since it maintains the marginal adaptation of indirect composite resin restorations, and improves the bond strength at the interface on buccal walls, which are the most critical regions for the long-term durability of these procedures.