Polycystic Ovary Syndrome (PCOS) is a frequent endocrine disease affecting women in their childbearing ages, often causing infertility and long-term metabolic problems. Early diagnosis is critical for successful management and better patient outcomes. This research proposes a machine learning prediction model based on a large dataset of 2,000 clinical, demographic, and hormonal data. Several classification models were attempted, and the Multilayer Perceptron (MLP) showed optimal performance for accuracy and F1 score. We were capable of offering worldwide and individual explanations by employing SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-Agnostic Explanations), which made the predictions more believable. The system is more user-friendly and accessible at clinics because of its cross-platform web application. The objective of this research is to diagnose PCOS in women at an earlier stage, utilizing results that are both interpretable and highly accurate. This will simplify and personalize healthcare services.