Objectives The aim of this split-mouth randomized controlled clinical trial was to compare the deviations of planned and placed implants placed by the assistance of a micron tracker-based dynamic navigation device or freehand methods. Material and methods A thermoplastic fiducial marker was adapted on the anterior teeth, and cone-beam computerized tomography was used for imaging. A minimum of one implant was planned for each side of the posterior maxilla, and the dynamic navigation device or freehand method was randomly used for surgical insertion. Deviations were measured by matching the planning data with a final CBCT image. Linear deviations (mm) between the planned and placed implants were the primary outcome. The results were analysed by generalized linear mixed models (p < .05). (NCT03471208). Results A total of 92 implants were placed to 32 volunteers, and 86 implants were included in the final analysis. For the linear deviations, mean of differences (Delta) was 0.72mm (Standard deviation (SD): 0.26); (95% Confidence interval (CI): 0.39-1.02) in the shoulder of the implants (p < .001) and 0.69mm (SD: 0.36); (95% CI: 0.19-1.19) in the tip of the implants (p < .001). For the angular deviations, Delta was 5.33 degrees (SD: 1.63); (95% CI: 7.17-3.48); (p < .001). Conclusions The navigation technique can be used to transfer virtual implant planning to the patient's jaw with increased accuracy.