Accurate age determination is fundamental in both forensic medicine and anthropology. Many methods that relate dental characteristics to adult age have been proposed, but there is still no simple and reliable method that does not damage the study material. The aim of this work was to propose a relevant and practical technique for determining age in adults that could be used in both living and deceased individuals. The sample was composed of 210 CT scans from individuals aged from 15 to 85 years old, with four healthy canines present in the mouth. The 840 canines were modelled using Mimics® 10.01 software. The pulp volume/total volume ratio ×100 was determined for each tooth. Seven mathematical models, corresponding to all possible real situations, were determined by the weighted least squares method and ranked in order of relative performance. The adequacy of the seven models to the data was very high with the regressions proposed (0.915 < R (2) adjusted < 0.964). Ranked in order of performance, the maxillary model was the most powerful of the seven models for age determination, followed by the 4 canines model, the 13 model and the 23 model.