Polycystic Ovary Syndrome (PCOS) is a prevalent endocrine disorder linked to infertility and metabolic complications. Traditional diagnostics are often costly and time-consuming, necessitating efficient automated screening tools. This paper proposes an AI-driven PCOS risk prediction and health report generation system utilizing the CatBoost machine learning algorithm. The model was trained on clinical and lifestyle parameters, including follicle counts, AMH levels, and menstrual cycles. Comparative analysis against XGBoost and LightGBM revealed that CatBoost achieved superior performance with 91% accuracy and an AUC of 0.95. The system is deployed via a Streamlit web application, featuring real-time risk visualization and automated professional PDF report generation. Key features like Follicle No. (R) were identified as primary predictors. This research provides a scalable, technology-driven solution for early PCOS awareness and preliminary clinical decision support, significantly enhancing accessible healthcare analytics for women.