OBJECTIVE The aim of this study was to evaluate the marginal adaptation of direct Class II sandwich restorations with resin-modified glass-ionomer cements and compomers in comparison to base and total bond restorations. For sandwich restorations with a triple cure resin-modified glass-ionomer cement the influence of different light curing techniques was also evaluated. METHODS Large butt-joint class II cavities with cervical margins 1 mm below the cemento-enamel junction were cut into 120 extracted human molars. The cavities (15 groups, n = 8) were filled using a sandwich, base or total bond technique with materials from five different manufacturers. A three-sited light curing technique was used in 13 groups. For the material combination Vitremer/Z100 two additional groups with a different wand positioning and a metal matrix were evaluated. After water storage for 21 days and thermocycling (2000x, 5-55 degrees C), replicas were quantitatively analysed in the SEM. Statistical analysis was performed with the Kruskal-Wallis test and the Bonferroni test at p < 0.05. RESULTS The marginal adaptation of vertical enamel margins was not dependent on the restorative technique. For margins in dentine, marginal adaptation was significantly better with the sandwich technique than with a base or total bond technique for all materials. There were no significant differences between the base and total bond technique. Overall, resin-modified glass-ionomer cements showed somewhat better results than compomers in sandwich restorations, though differences were not significant for some criteria. Vitremer/Z100 sandwich restorations applied with a metal matrix showed the highest mean percentage of excellent margins of all experimental groups. CONCLUSIONS Both resin-modified glass-ionomer cements and compomers can improve the marginal quality when used in a sandwich technique. Further research is necessary to determine the ideal material combination for sandwich restorations.