APurpose : the study was designed to evaluate the trueness of two premolar endocrowns preparation design with different virtual impression techniques Material and methods: Two premolar teeth were endodontically treated, prepared using a CNC machine. Tooth A was prepared to have a flat occlusal surface, while tooth B was prepared to have a 45o cusp anatomy inclination. For each tooth, a total of 40 different impressions were obtained. They were divided into 4 groups (n=10) as follows: Group 1; direct scan using Medit i500, group 2; direct scan using trios 3shape, group 3; Putty and light impression were taken and impression was scanned using Identica Hybrid, group 4; Impressions were poured in gypsum and tooth model was scanned using identica hybrid. Both CAD reference models and all (STL) files (n=80) were loaded into a 3D reverse engineering software (Geomagic QualifyTM 2012, Geomagic, Morrisville, USA). By using this method, for each superimposition, the deviation at each measurement point was recorded as a root mean square. Results : There statistically significant differences between different groups. Anatomical reductions showed statistically significant higher trueness when compared to flat occlusal reduction. As for impression method, no statistically significant differences between direct Trios 3Shape scanning and impression scanning. However, both showed higher trueness when compared to direct Medit i500 scanning and cast scanning. Conclusion : Anatomical preparation can be reproduced more accurately than flat preparation. Scanning direct impression produce more accurate virtual impression.