Whereas cranial neuronavigation is widely accepted as a helpful tool, larger series of the in vivo application of spinal neuronavigation do not exist. In the following we report our 4-year experience with spinal navigation in 75 consecutive cases for dorsal transpedicular screw placement. Seventy-five patients were planned for operation employing anatomical reference points defined on a 2-mm high resolution CT. We used single vertebra registration and surface matching. With the above methods, the mean registration deviation ranged from 0.18 mm (cervical spine) to 0.31 mm (lumbar spine). All our screws in the upper cervical spine were navigated correctly (17 patients), thus improving markedly the surgical outcome. The results were not as promising in the lumbar area. In only 84% was navigation reliable. The reason was the lack of a practicable tracking tool. Spinal neuronavigation based on anatomical reference points is able to improve the results in transpedicular screwing, especially in the cervical spine. The lack of a practicable tracking tool still hinders its use in routine clinical application.