BACKGROUND This study aimed to evaluate the effect of different percentages of ethanol solvent of an experimental methacrylate-based dentin bonding agent containing polyhedral oligomeric silsesquioxanes (POSS) on the microleakage of resin composite restorations. MATERIALS AND METHODS In this experimental study, 42 extracted human premolar teeth used and 84 standard Class V cavities were prepared on the buccal and lingual surfaces of the teeth. The teeth were divided into 6 groups of 7. Experimental bonding agents with different percentages of solvent were used in 5 groups and Single Bond® as a control. The teeth were restored with resin composite and subjected to thermal cycling test. Teeth were then immersed in a solution of 2% basic fuchsine dye for 24 h and sectioned buccolingually and scored using stereomicroscope with ×32 magnification. Microleakage data were analyzed using the Kruskal-Wallis, Mann-Whitney U, and Wilcoxon tests. RESULTS There were significant differences between the microleakage enamel margins (P = 0.036) and dentinal margins (P = 0.008) in all the groups. These significant differences were seen between the control group and groups containing 46 wt% solvent (P = 0.011), 46 wt% and 31 wt% solvent in dentinal (P = 0.027), 31 wt% and 0 wt% in enamel (P = 0.021), also 0 wt% and control in enamel (P = 0.039), and dentinal margins microleakage (P = 0.004). The microleakage in dentinal margins was higher than enamel margins (P < 0.001). In the groups with 46 wt% solvent (P = 0.103), 0 wt% (P = 0.122), and control group (P = 0.096), however, this difference was not significant. CONCLUSION The adhesive containing 31 wt% solvent showed the least marginal microleakage, presence of POSS filler may also result in the reduction of microleakage.