The achievable accuracy is a decisive parameter for the comparison of direct intraoral digitization with the conventional impression. The objective of the study was therefore to compare the accuracy of the reproduction of a model situation by intraoral digitization vs. the conventional procedure consisting of impression taking, model production, and extraoral digitization. Proceeding from a die model with a prepared tooth 16, the reference data set of the teeth 15, 16 and 17 was produced with an established procedure by means of extraoral digitization. For the simulated intraoral data acquisition of the master model (Cerec 3D camera, Sirona, Bensheim), the camera was fastened on a stand for the measurement and the teeth digitized seven times each in defined views (occlusal, and in each case inclined by 20°, from the mesio-proximal, disto-proximal, vestibular and oral aspect). Matching was automated (comparative data sets B1-B5). A clinically perfect one-step putty-and-wash impression was taken from the starting model. The model produced under defined conditions was digitized extraorally five times (digi- SCAN, comparative data sets C1-C5). The data sets B1-B5 and C1-C5 were assigned to the reference data set by means of best-fit matching and the root of the mean quadratic deviation (RMS; root mean square) calculated. The deviations were visualized, and mean positive, negative and absolute deviations calculated. The mean RMS was 27.9 µm (B1-B5) or 18.8 µm (C1-C5). The mean deviations for the prepared tooth were 18 µm/-17 µm (B1-B5) and 9 µm /-9 µm (C1-C5). For tooth 15, the mean deviations were 22 µm/-19 µm (B1-B5) and 15 µm/ -16 µm (C1-C5). The intraoral method showed good results with deviations from the CAD starting model of approx. 17 µm, related to the prepared tooth 16. On the whole, in this in-vitro study, extraoral digitization with impression taking and model production showed higher accuracy than intraoral digitization. Since the inaccuracies in the conventional impression under real clinical conditions may be higher than the values determined above, a comparison under clinical conditions should be performed subsequently. Palabras clave: intraoral digitization, matching, accuracy, one-step puttyand- wash impression