This study investigates the use of machine learning algorithms for classifying various grasp postures based on data collected from a data glove equipped with resistive flex sensors. The glove measures the flexion and extension of the fingers, translating hand movements into electrical signals that are used to train and evaluate five machine learning models: Support Vector Machine (SVM), Decision Tree, Random Forest, k-Nearest Neighbors (KNN), and Logistic Regression. The models were assessed on their accuracy, precision, recall, and computational efficiency. Among the algorithms, KNN and SVM demonstrated the highest classification accuracy, particularly for grasp types such as Cylindrical Grip, Diagonal Volar Grip, and Lateral Pinch. However, the Hook Grip posed a challenge for several models, indicating the need for further improvement in feature extraction and model optimization. While Random Forest provided robust accuracy, its higher computational time made it less suitable for real-time applications. Conversely, Decision Trees offered faster computation but at the expense of lower accuracy. Logistic Regression displayed the weakest performance overall. The results underscore the importance of selecting the appropriate machine learning model based on the specific requirements of the grasp classification task. This research contributes to the development of more accurate and efficient gesture recognition systems, with significant implications for applications in robotics, neurorehabilitation, and assistive technologies where precise and real-time grasp recognition is critical.