Purpose Scaphoid fractures are commonly fixed with headless cannulated screws positioned centrally in the scaphoid. Judgement of central placement of the screw may be difficult. We generated a central zone using computer analysis of 3D reconstructions of computed tomography (CT) images. As long as the screw axis is completely contained within this central zone, the screw would be considered as centrally placed. Methods Thirty cases of 3D CT reconstructions of normal scaphoids in a computerised operation planning and simulation system (Vxwork software) were obtained. The central zone was established after some distance shrinkage of the original scaphoid surface reconstruction model using the function "erode" in the software. The shape of the central zone was evaluated, and the width of the central zone in the proximal pole, waist portion and distal pole was measured. We also established the long axis of the scaphoid to see whether it stays in the central zone. Results All central zones could be divided into distal, waist and proximal portions according to the corresponding irregular shape of the scaphoid. As the geometry of the central zone was so irregular and its width very narrow, it was possible to completely contain the screw axis either in the proximal portion alone, waist alone or distal central zone alone. Conclusions Establishing the central zone of scaphoid 3D CT images provided a baseline for discussion of central placement of a scaphoid screw. The geometry of the scaphoid central zone determined that the screw could hardly be inserted through entire scaphoid central area during surgery.