Automated classification of skin fungal infections is crucial for enhancing diagnostic accuracy. This study explores machine learning methods using a dataset of 1,158 images, employing preprocessing and feature extraction techniques (texture, structural, and statistical), followed by feature selection methods: PCA, p-test, and top-no We assessed traditional classifiers (Support Vector Machine, Logistic Regression, K-Nearest Neighbors) alongside ensemble methods (Random Forest, XGBoost, Voting, Bagging, Boosting). The p-test with Random Forest yielded the highest accuracy (88.41%), followed closely by XGBoost (87.98%). This research demonstrates the effectiveness of integrating feature selection techniques with machine learning models to improve the classification of skin fungal infections and enhance dermatological diagnostics.