Objective: To characterize the three-dimensional mandibular morphology of adults with skeletal Class II malocclusion and mandibular deviation using cone-beam computed tomography (CBCT), identify independent morphologic indicators, and construct a combined discrimination model. Methods: A total of 126 adults with skeletal Class II malocclusion and mandibular deviation were retrospectively included as the deviation group, and 118 adults with normal occlusion were selected as the normal group. Three-dimensional mandibular reconstruction and measurements were performed using Mimics 20.0. Paired and independent-samples t tests were used for group comparisons. Variables showing between-group differences were entered into univariable logistic regression; multivariable logistic regression was then performed after correlation and variance inflation factor assessment. Receiver operating characteristic (ROC) curves were used to determine the area under the curve (AUC), optimal cut-off, sensitivity, and specificity. Internal validation was performed with 1,000 bootstrap resamples. Results: No significant left-right differences were observed in the normal group (all P>0.05). On the deviated side, condylar length, mandibular body unit length, condylar width, ramus height, mandibular body length, and three volume measurements were lower than those on the contralateral side (all P<0.001). Compared with the normal group, nine morphologic variables were lower in the deviation group, whereas mandibular angle length and chin length did not differ significantly. Condylar length, mandibular body unit length, condylar width, and ramus height remained independently associated with skeletal Class II malocclusion with mandibular deviation in the multivariable model. The combined model yielded an AUC of 0.925 (95% CI, 0.893-0.958), with an optimal probability cut-off of 0.46, sensitivity of 88.10%, and specificity of 84.75%. The bootstrap-corrected AUC was 0.920. Conclusion: Mandibular asymmetry in adults with skeletal Class II malocclusion and mandibular deviation is mainly concentrated in the condyle, ramus, and part of the mandibular body. A combined model based on condylar length, mandibular body unit length, condylar width, and ramus height provides useful quantitative support for three-dimensional CBCT assessment.