OBJECTIVE The purpose of this study was to analyze the sealing ability of different temporary endodontic materials over a 6-week period using a glucose penetration model. METHODS Standardized holes were formed on 48 dentin discs from human premolars. The thicknesses of the specimens were distributed evenly to 2 mm, 3 mm and 4 mm. Prepared dentin specimens were randomly assigned into six groups (n = 7) and the holes in the dentin specimens were filled with two kinds of temporary filling materials as per the manufacturers' instructions as follows: Caviton (GC Corporation, Tokyo, Japan) 2 mm, 3 mm, 4 mm and IRM (Dentsply International Inc., Milford, DE) 2 mm, 3 mm, 4 mm. The remaining specimens were used as positive and negative controls and all specimens underwent thermocycling (1000; 5-55°C). The sealing ability of all samples was evaluated using the leakage model for glucose. The samples were analyzed by a spectrophotometer in quantitative glucose microleakage test over a period of 6 weeks. As a statistical inference, a mixed effect analysis was applied to analyze serial measurements over time. RESULTS The Caviton groups showed less glucose penetration in comparison with the IRM groups. The Caviton 4 mm group demonstrated relatively low glucose leakage over the test period. High glucose leakage was detected throughout the test period in all IRM groups. The glucose leakage level increased after 1 week in the Caviton 2 mm group and after 4 weeks in the Caviton 3 mm and 4 mm groups (p < 0.05). CONCLUSIONS Caviton had better sealing ability than IRM in the glucose penetration model during 6 weeks. Temporary filling of Caviton to at least 3 mm in thickness is necessary and temporary filling periods should not exceed 4 weeks.