This study aimed to use deep learning to segment the mandible and identify three-dimensional (3D) anatomical landmarks from cone-beam computed tomography (CBCT) images, and to construct the mandibular median sagittal plane (MMSP) to assess mandibular symmetry. A total of 400 participants were randomly divided into a training group (n=360) and a validation group (n=40), with 50 morphologically normal individuals used as the test group. The PointRend deep learning mechanism was used to segment the mandible and identify 27 anatomic landmarks via PoseNet. The 3D coordinates of 5 central landmarks and 2 pairs of side landmarks were obtained for the test group. The template mapping technique was used to screen 35 combinations of 3 midline landmarks, and the asymmetry index (AI) was calculated for each of the 35 mirror planes. The B-Gn-F plane was found to have the smallest AI (1.6), and it was determined to be the most accurate MMSP. The study concluded that deep learning can automatically segment the mandible, identify anatomic landmarks, and provide a simple and accurate method for clinical judgment of mandibular symmetry. The B-Gn-F plane was identified as the optimal MMSP for clinical application due to its symmetry and stability.This study aimed to use deep learning to segment the mandible and identify three-dimensional (3D) anatomical landmarks from cone-beam computed tomography (CBCT) images, and to construct the mandibular median sagittal plane (MMSP) to assess mandibular symmetry. A total of 400 participants were randomly divided into a training group (n=360) and a validation group (n=40), with 50 morphologically normal individuals used as the test group. The PointRend deep learning mechanism was used to segment the mandible and identify 27 anatomic landmarks via PoseNet. The 3D coordinates of 5 central landmarks and 2 pairs of side landmarks were obtained for the test group. The template mapping technique was used to screen 35 combinations of 3 midline landmarks, and the asymmetry index (AI) was calculated for each of the 35 mirror planes. The B-Gn-F plane was found to have the smallest AI (1.6), and it was determined to be the most accurate MMSP. The study concluded that deep learning can automatically segment the mandible, identify anatomic landmarks, and provide a simple and accurate method for clinical judgment of mandibular symmetry. The B-Gn-F plane was identified as the optimal MMSP for clinical application due to its symmetry and stability.