The pelvis and skull are the regions of the human skeleton that most clearly show sexual dimorphism, with the pelvis being superior to the skull for sex estimation owing to reproductive and hormonal factors. However, as many skeletons are found incomplete, it is important to be able to determine sex from analyzing the skull as well as the pelvis. This study was aimed at validating and evaluating the effectiveness of four morphometric methods for sex estimation using the human skull. One hundred human Brazilian skulls belonging to identified but unclaimed skeletons were used after all legal and administrative requirements for exhumation were satisfied. Four dimensions were analyzed: the facial triangle; the bimastoid triangle; the mastoid triangle; and finally, the occipital triangle. These dimensions were used to calculate the triangle areas. Descriptive analysis was used to verify the mean, standard deviation, and confidence interval of all variables and discriminant analysis with cross-validation was used to verify the percentage of correctness of the sex estimation methods. Individual analysis of the areas of the facial, right mastoid, left mastoid, bimastoid, and occipital triangles, and sum of the mastoid triangle areas, produced average accuracies of 63.0%, 81.8%, 77.8%, 71.4%, 64.0%, and 80.8%, respectively. Combinations of triangles resulted in higher mean accuracy percentages, with only 75.6% accuracy recorded when the left and right mastoid triangle areas were removed. All average values of analyzed triangles were higher in males than in females. Analyses of the cranial triangles were useful for sex estimation in the population studied.