BACKGROUND Despite the growing tendency toward tooth-colored restorations in dentistry, polymerization shrinkage and subsequent marginal microleakage remains a problem. The aim of this in vitro study was to compare microleakage between silorane-based and methacrylate-based composite resins at different time intervals and with different restorative techniques. METHODS In this in vitro study, 108 sound extracted human molar teeth were used. Mesial and distal proximal class II boxes with dimensions of 1.5 mm depth and 4 mm width were prepared. The gingival margins of all cavities were 1 mm below the cement enamel junction. The teeth were randomly divided into three groups based on test materials. In the first group, the teeth were restored by a nanocomposite (Filtek Z350XT, 3MESPE) and SE Bond adhesive (Kuraray, Japan), in the second group, the teeth were restored with a silorane-based (Filtek P90, 3MESPE) and Filtek P90 Adhesive (3M ESPE, USA) and in the third group, the teeth were restored with a microhybrid posterior composite resin (Filtek P60, 3MESPE) and SE Bond adhesive (Kuraray, Japan). Half of the proximal cavities in each of these three groups were restored in two horizontal layers and the other half in four horizontal layers. After a period of aging (24-h, 3-month and 6-month) in water and then application of 500 thermal cycles, the teeth were immersed for 24-h in 0.5% fuchsin and evaluated under a stereomicroscope at ×36 magnification to evaluate leakage in gingival margin. Data was statistically analyzed using Kruskal-Wallis and Mann-Whitney U-tests. P ≤ 0.05 was considered as significant. RESULTS In Z350XT statistically significant differences were observed in microleakage in comparison of 24-h and 6-month intervals (P = 0.01) that was higher in 6-month. Comparison of microleakage in P90 and P60 composite resins was also statistically significant and was less in P90. Microleakage was not significantly different between P90 and Z350XT at 24-h. However, this difference was significant at 3-month and 6-month intervals. Differences in microleakage of P60 and Z-350XT composite resins were not statistically significant in all intervals (P = 0.38). P90 showed the lowest microleakage during storage in water. Z350XT had microleakage similar to P90 within 24-h, but after 6-month of storage in water, it showed the highest microleakage among all the groups. The number of layers (2 layers vs. 4 layers) did not result in any differences in microleakage scores of the composite resins (P = 0.42). CONCLUSIONS Water storage times did not result in any significant effect on microleakage of P90 and P60.