MRI based scanning can record hard tissue behind soft tissue with a clear demarcation between them. Hence, this type of scanning may be considered a potential alternative for making digital impressions of tooth structure, particularly in the case of sub–gingival tooth preparations. This study aimed to analyze the accuracy of MRI based impressions made from teeth with sub–gingival preparation compared with impressions obtained through digital optical scanning. Twelve extracted human teeth (Maxillary pre–molars) were prepared for crowns with a chamfer finish line. These samples were divided into four groups with three samples each. Samples under group 1 and group 2 were devoid of gingival coverage. whereas, samples under group 3 and group 4 were covered with porcine gingiva (thickness ranged between 2mm and 4mm). Samples under groups 1 and 3 were subjected to optical scanning using CS2 extra–oral scanner (STRAUMANN), while samples under group 2 and 4 were subjected to MRI based scanning (SIEMENS). The dataset obtained from both scanners were converted to STL files. These STL files were utilized to 3D print twelve all–ceramic crowns. The accuracy of crowns was compared to evaluate the marginal and internal fit. The values were statistically analyzed using the student t–test. The digital impression of sub–gingivally prepared teeth was achievable using MRI scanners. They were able to scan teeth with accuracy better than that of optical scanners when samples were covered with gingiva.