Parkinson’s disease (PD) is a neurodegenerative condition that severely impairs motor function and reduces quality of life. Effective disease management depends on early and precise diagnosis. Recent advancements in artificial intelligence have significantly enhanced the accuracy of PD diagnosis, particularly through the analysis of speech data. In this paper, we propose a machine learning algorithm based on voice features to classify PD and Healthy Controls (HC). We employed various machine learning classifiers, including Random Forest (RF), Logistic Regression, KNN, Support Vector Machine (SVM), and XGBoost. After hyperparameter tuning, XGBoost achieved the highest accuracy at 96.67%. KNN, RF, and SVM followed with accuracies of 94.87%, and Logistic Regression (LR) recorded the lowest accuracy at 89.74%. Our results demonstrate that effective hyperparameter tuning played a crucial role in achieving higher accuracy compared to similar studies using the same classifiers.