One of the known problems in forensic medicine is identification and sex determination which is in need of knowledge and expertise in many fields. In generating the biological profile of human remains, sex is a factor of primary significance. In cases where intact skull is not found, the mandible may play a vital role in sex determination as it is the most dimorphic bone of skull. It is well known fact that skeletal structures vary among different populations and that each therefore needs its own specific standards of assessment to optimize the accuracy of identification. The current study focuses on sexual identification from mandibular bone in Saudi sample population. The study comprised 80 Saudi persons (40 males and 40 females) their age ranged from (20-55 years) visited the Radiological Department in Elite Private Hospital in Riyadh City in Saudi Arabia Kingdom. All cases were submitted to three-dimensional (3D) Computed Tomography.  All the participants' skulls were examined and six mandibular measurements were described, four of them (Mandible angle, Mandibular base length, Ramus length and Minimum ramus breadth) were measured from the lateral reconstruction 3D CT image and two measurements (Bigonial breadth and Bicondylar breadth) from axial posterior reconstruction 3D CT image. In all the studied measurements, male mandibles were found to have significantly higher values compared to female mandibles. By comparison the mean values of the studied six mandibular measurements between males and females, the results revealed that there are four mandibular measurements considered as final predictors of sex determination which are (mandible angle, mandibular base length, ramus length and the bicondylar length). The overall predictive accuracy of this prediction model was 84.95% for the whole studied group and it could correctly identify males in 84.7% and females in 85.2% cases. The study concluded that sex can be determined by using mandible dimensions with relatively high accuracy and mandible is a suitable bone for forensic analysis.