Purpose: To evaluate the accuracy of available bone width, height, and length measurements on preplanned implant sites using cone beam computed tomography (CBCT) images scanned at different angulations of the mandible. Materials and Methods: Standard cylindrical holes were prepared on six dry human mandibles and filled with warm gutta-percha to create spherical markers for measurements of available bone width, height, and length. Mandibles were first scanned with a CBCT device in the ideal position with the occlusal plane parallel to the horizontal plane. Then, images of the mandibles were obtained in rotation, tilt, flexion, and extension positions using 5-and 10-degree angulations. Measurements were done on a total of 54 images. Original dimensions of the available bone for planned implant sites were measured with a digital caliper on dry mandibles as the gold standard. The absolute values of the differences between each measurement and the gold standard were obtained for measurement errors. Repeated-measures analysis of variance and Dunnett's multiple comparisons test were used for comparisons (P=.05). Intraobserver and interobserver agreement was calculated using intraclass correlation coefficient (ICC). Results: ICC was excellent for both intraobserver and interobserver reproducibility. No significant difference was found between length and height measurements in the ideal position and in rotation, tilt, flexion, and extension movements of mandibles at two different angulations (P > .05). Width measurements revealed a significant difference among the ideal position and measurements at the 10-degree flexion, 10-degree extension, 10-degree rotation, and 10-degree tilted mandibular positions (P < .05). Conclusion: The position of the occlusal plane with respect to the floor during the CBCT scan may have a clinically significant effect on dental implant site dimensions.