This work advances privacy-preserving reproductive health monitoring by combining contactless biosensing technology with federated learning (FL) to track menstrual cycles. Skin-contact devices or manual inputs can be inaccurate, obtrusive, and privacy-compromising, especially when sensitive health data is centrally kept. These restrictions might make reproductive health management difficult, especially for irregular cyclers. A contactless sensing technology that continually measures important physiological signals like heart rate and breathing without skin contact addresses these problems. We integrate radar-based physiological sensing, photoplethysmography (PPG), and LiDAR monitoring. Our approach uses FL to keep sensitive health data on individual devices, lowering the danger of data breaches and improving user privacy. The technology uses multimodal AI models to analyze physiological information to enhance menstrual cycle forecasts and adaptability. Users with irregular patterns benefit from this. The result is a discreet, accurate, and personalized menstrual health tracking system. This research combines cutting-edge contactless sensing with privacy-focused AI to change reproductive health management and research.