Compomer is composed of matrix and filler; matrix is made of the combination of resins and polycarboxylic molecules that are light-cured, and a filler is a glass component which is capable of ion-release. The resin content of compomers produces polymerization shrinkage which can adversely affect marginal adaptation. Pretreatment is a fundamental step which is treated with conditioner or primer in the use of these materials. Microleakage of restorative materials has been investigated mostly by dye penetration method. Dye penetration method was not quantitative and not measured repeatedly. Fluid filtration method, introduced and developed by Pashley's group, has been extensively used for 20 years for research purpose to understand the physiology of dentin, as well as the effects of various restorative treatments on dentin permeability. It permits quantitative, nondestructive measurment of microleakage in a longitudinal manner. The purpose of this study was to evaluate the change of dentin permeability according to the process of compomer restoration. In this study, ClV cavities were prepared on buccal surface of thirty extracted human molars. The prepared cavities were etched by 37% phosphoric acid. The experimental teeth were randomly divided into three groups. Each group was treated with following materials; Group 1 : Prime & Bond NT/Dyract AP, Group2 : Single Bond/F2000 compomer, Group 3 : Syntac Single Component/Compoglass. The bonding agent and compomer were applied for each group following manufacturers information. Dentin permeability of each group was measured at each process by fluid filtration method; Step 1 : preparation(smear layer), Step 2 : etching(smear layer removal), Step 3 : applying the bonding agent, Step 4 : filling the compomer. Dentin permeability was expressed by hydraulic conductance(microl min(-1)cmH2O(-1)). The data were analysed statistically using One-way ANOVA and Sheffe's method. The results were as follows : 1. Dentin permeability differences between each process were significant except between step 1 and step 2(p<0.01). 2. Dentin permeability after removal of smear layer was highly increased(p<0.01). 3. In most case, decrease of dentin permeability was obtained by applying bonding agent(p<0.01). 4. Dentin permeability differences among the experimental groups were not significant(p>0.05). 5. None of compomers used in this study showed perfect seal at the interface.