Objective: To evaluate the three-dimensional (3D) reconstruction accuracy of the intercuspal occlusion (ICO) of the dental casts, by the dental articulator position method, and provide a refe-rence for clinical application. Methods: The standard dental casts in ICO were mounted on average values articulator, and five pairs of milling resin cylinders were respectively attached to the base of both the casts. 100 μm articulating paper and occlusal record silicone rubber were used to detect the occlusal contact number between the posterior teeth of casts mounted on articulator in ICO. The occlusal contact numbers N_A detected by the two methods were calculated simultaneously, as the reference. After the upper and lower casts were scanned separately, and the buccal data of casts in ICO were scanned with the aid of the dental articulator position, registration was carried out utilizing the registration software. Then the digital casts mounted in ICO as well as the buccal occlusal data were saved in standard tessellation language (STL) format. Geomagic Studio 2013 software was used to analyze the contact number N_S between digital upper and lower casts by the “deviation analysis” function. The differences were compared between N_S and N_A, to evaluate the sensitivity and positive predict value (PPV) of the model scanner to reproduce the occlusal contact with the aid of dental articulator position. The distance D_R between the centers of the circles at the top surface of the upper and lower corresponding cylinders was obtained by the three-coordinate measuring system Faro Edge, as the reference value. The Geomagic Studio 2013 software was used to construct the cylinders of digital casts and the distance D_M between the centers of the circles at the top surface of the corresponding cylinders were measured, as the test value. The one-sample T test was used to analyze the variable differences between D_M and D_R. Results: The PPV of reproducing the occlusal contact point was 0.76 and sensitivity was 0.81. The distance error of the opposite cylinder was (0.232±0.089) mm. There was no statistical difference between the feature points 5-5’, while there were statistical differences between the other feature points. Conclusion: By the dental articulator position method, the model scanner reproduces the occlusal contact point with high sensitivity and PPV, and that meets clinical needs. Meanwhile, the distance between the feature points is greater than the reference value, which will lead to occlusal disturbance, and require clinical grinding.