Purpose: To accurately accomplish the drilling of an implant socket, the use of image-guided navigation has become an option. The aim of this study was to evaluate the 3-dimensional (3D) accuracy of navigation-guided drilled holes. Materials and Methods: Laboratory accuracy measurements were obtained on an acrylic resin model with standardized target holes drilled by a computerized numerical control machine. The model was scanned by a multislice computerized tomography scanner and registered with fiducial marker-based algorithms. Navigated drillings were performed using an optical navigation system based on passive marker technology. Coordinates of drilled holes were determined by a 3D-digitizer probe, and accuracy was assessed for all 5 degrees of freedom using a computer-aided design system (Pro/Engineer). Results: A total of 240 drillings were evaluated. Mean registration error was 0.86 mm (SD 0.25 mm). Target point deviation between preplanned and actual drill starting point was 0.95 mm (SD 0.25 mm). The deviation in terms of full length was 0.97 mm (SD 0.34 mm), and mean angular deviation on the coronal and sagittal planes was 1.35 degrees (SD 0.42 degrees). Discussion: The accuracy of image-guided navigation depends on imaging modalities, patient-to-image registration procedures, and instrument tracking. The technical accuracy and the navigation procedure, as evaluated in the study presented, seem to be of minor influence. Conclusion: The data obtained by this in vitro study demonstrate that the accuracy of navigation-based drilling may be sufficient for clinical practice, particularly in terms of the transferability of preplanned trajectories. However, in vivo clinical trials need to be performed to evaluate the clinical accuracy and treatment quality of navigation-guided interventions.