Falls are a leading cause of injury and hospitalization among the elderly, most of the existing solutions remain reactive; alerts are sent only after a fall has occurred. This paper proposes SAFE-PRO (Smart Assistive Framework for Falls in the Elderly – Proactive Risk observation), a wearable system designed to anticipate fall risk and provide real-time preventive alerts. SAFE-PRO uses multiple sensors, such as tri-axial accelerometer, gyroscope, and heart rate sensor to detect gait anomalies, motion instability, and physiological signs of distress. A light-weight machine learning model is deployed on the device itself, enabling, low-latency inference without dependence on cloud infrastructure. If instability is detected, the system sends a haptic audio alert, giving the user a chance to correct themselves or get help before a fall happens. The proposed architecture represents a shift from traditional fall detection towards proactive fall risk ideology. The proposed design is expected to support real-time, on-device prediction with minimal resource overhead, indicating SAFE-PRO’s potential suitability for continuous use in elderly care.