Gender classification is crucial in fields like security, biometrics, and sports science. Traditional methods relying on physical characteristics, such as genital shape, have become less reliable due to modern challenges like sex reassignment surgeries. This study proposes a more reliable approach using anthropometric data from the ANSUR II dataset, applying machine learning algorithms such as Logistic Regression, Neural Networks, and SVMs. Our results show that high classification accuracy (up to 100%) can be achieved using only ten selected features. This is a significant improvement over previous methods, which required more features and complex models. Notably, our approach outperformed the latest studies, which used 25 features and achieved lower accuracy. The alignment of key measurements, like biacromial breadth, with those identified in recent research further validates our feature selection process. This research contributes to the development of more accurate and efficient gender classification systems that can adapt to the evolving challenges posed by modern societal changes, ensuring fairness and reliability in applications ranging from competitive sports to security screening.