The use of flowable composite resins as liners in class II packable composite restoration has been suggested by some manufacturers. However, the contributions of this technique are unproven. The purpose of this study was to compare the gingival microleakage in class II packable composite restorations with or without the use of flowable composite resins as liners. Slot cavities were prepared on both proximals of 80 extracted human molars and randomly assigned to 8 groups of 20 each. The gingival margins were located at 1mm above CEJ in 80 cavities (group1-4) and 1mm below CEJ in 80 cavities (group5-8). The prepared teeth were mounted in the customized tray with adjacent teeth to simulate clinical conditions and metallic matrix band (Sectional matrix) and wooden wedges were applied. After acid etching and application of Single Bond, each group was restored with the following materials using incremental placement technique: Group 1,5 (Filtek P60), group 2, 3, 4 and group 6, 7, 8 (AeliteFlo, TetricFlow, Revolution/ Filtek P60). All specimens were thermocycled 500 times between 5$^{circ}C$ and 55$^{circ}C$ with 1 mimute dwell time, immersed 2% methylene blue dye for 24 hours and then rinsed with tab water. The specimens were embedded in clear resin and sectioned longitudinally through the center of restoration with a low speed diamond saw. Dye penetration at gingival margin was viewed at 20 magnification and analyzed on a scale of 0 to 4. Kruscal-Wallis One way analysis and Mann-Whitney Rank sum test were used to analyze the results. The results of this study were as follows. 1. The leakage values seen at the enamel margin were significantly lower than those seen at the dentin margin(P<0.05). 2. On the enamel margin, packable composite resins with flowable liners showed lower leakage than those without flowable liners, but there were no significant differences among the four groups(P>0.05). 3. On the dentin margin, four groups demonstrated moderate to severe leakage, and there were no significant differences in leakage values(P>0.05).