One of the aims of the present investigation was to assess three-dimensionally the anteroposterior discrepancy of dental bases using a noninvasive direct procedure. A second aim was to verify the relationship of three-dimensional soft-tissue measurement to the well-established two-dimensional cephalometric assessments of anteroposterior discrepancy. Dental and facial landmarks were directly digitized on 20 orthodontic and maxillofacial surgery patients aged 8 to 26 years using an electromagnetic three-dimensional computerized digitizer. The anteroposterior maxillomandibular discrepancy was measured by calculating the linear distances between the projections of subnasal and sublabial landmarks on the occlusal plane, subnasal and sublabial landmarks on Camper's plane, and insertion of maxillary and mandibular median labial frenula on the occlusal plane. From lateral cephalograms of the same patients, the following measurements were obtained: subspinale point-nasion-supramentale point (ANB) angle; corrected ANB angle that compensates for the position of the maxilla and rotation of the mandible relative to the cranial base; Wits appraisal; MM-Wits, linear distance between the projections of points A and B on the bisector of the palatal plane to mandibular plane angle; and soft-tissue Wits, linear distance between the projections of soft-tissue points A and B on the bisecting occlusal plane. The best two-dimensional vs three-dimensional linear regression (r = 0.91) was found between Wits appraisal and the linear distances between the projections of maxillary and mandibular median labial frenula on the occlusal plane (Wits = -1.05 x 3D measurement - 3.75). The three-dimensional evaluation of the sagittal discrepancy of the jaws directly performed in vivo may allow a more complete analysis of a patient's soft-tissue drape together with the underlying hard-tissue structure.