Objective: To evaluate the morphology accuracy of direct occlusal veneer using two types of modified stamp-technique, comparing the results of two types of stamp and different composite resin. Methods: Model scanner was used to get the original data from the standard resin teeth in plastic model. Two types of stamps were made: solid silicon stamp and transparent silicon stamp. In the study, 54 resin tooth were randomly divided into 9 groups: Groups 1-8 were restored by direct composite resin, using general and bulk-filled composite resin in occlusal veneer (1 mm or 2 mm) with two types of stamp-technique (the solid silicon, and the transparent silicon), and the control group was restored by indirect way using CAD/CAM composite resin restoration by biocopy technique to mimic the original shape of the teeth. After the resin teeth were prepared for occlusal veneer, the direct and indirect composite resins were restored. After the restoration was finished, the data were obtained again by the same model scanner in the plastic model. 3D comparisons were made using the original and restoration data, the average difference and the root mean square of difference (RMS) were recoded from the software. The RMS of all the groups were analyzed using one way ANOVA (α=0.05). Results: The RMS of occlusal surface in solid silicon stamp group was (0.136±0.031) mm, in transparent silicon group was (0.130±0.024) mm, and in control group was (0.130±0.009) mm. There were no significant difference. While the RMS of buccal/lingual surface in solid silicon stamp group was the smallest [(0.135±0.020) mm, P<0.05], and in transparent silicon group it was (0.147±0.021) mm, and control group (0.153±0.014) mm. The general composite resin using the two types of stamp had the largest occlusal RMS in 2 mm occlusal veneer, which were significantly larger than control group (P<0.05). Conclusion: The direct occlusal veneer using two types of modified “stamp-technique” had relatively good morphology accuracy. Direct occlusal veneer using bulk-filled composite resin made with two types of stamp technique had even more accurate morphology than using general composite resin. The solid silicon stamp had a better morphology accuracy in buccal/lingual surface than the transparent silicon stamp.