This paper introduces a radiomics approach for legal age classification based on cone-beam computed tomography (CBCT) images of the mandibular condylar head. The study aims to identify age-related radiomics features and develop an age classification model. CBCT images from 85 subjects were analyzed, and 127 radiomics features were extracted using AVIEW software. The random forest method was employed to select the most important features and build age classification models for four legal age groups. The 21-year age classification model achieved the highest accuracy of 91.18%, demonstrating the potential of radiomics features as imaging biomarkers for age estimation.