To evaluate the marginal leakage in Class V cavity this study has been undertaken to compare the sandwich technique with the conventional method, and find out the effect of lining material, lining method, and polishing time on sandwich technique. Ninty extracted teeth were divided into eight test groups and a control group, and were prepared with a buccal Class V cavity. Four test groups were lined with Dentin Cement whereas the other four test groups were lined with Vitrabond. Half of the either group were lined 0.5mm short of the carvosurface margin and the rest were lined completely to the carvosurface margin. The four subgroups were further divided into specimens which were polished immediately and 24 hours after resin filling. The polished specimens were immersed in $37^{circ}C$, 0.5% methylene blue solution for 24 hous after thermocycling at $5^{circ}C$ and $55^{circ}C$, 200 times and buccolingually sectioned. The sectioned specimens were examined dye penetration under the light microscope. The following results were obtained, 1. At the enamel margin, the conventional method showed a lower microleakage than the sandwich technique. The difference between the control and Vitrabond group was statistically significant(P<0.05), but no difference between the control and Dentin Cement group, and between the lining materials was observed. 2. At the dentinal margin, the sandwich technique showed a significant lower amount of microleakage (P<0.05), but there was no significant difference between the lining materials. 3. Regardless of the lining material, lining method, and polishing time used, values of microleakage were significantly higher at the dentinal margin compared to the enamel margin(P<0.05). 4. In specimens till the cavosurface margin, microleakage at the dentinal margin was less with the light-cured base than with the chemically-cured base, but there was no siginificantly difference between the lining materials regarding the lining method and polishing time. 5. The lining material, lining method, and polishing time did not affect the amount of micro leakage in the sandwich technique(P<0.05).