A highly adaptive and proactive safety system design scheme integrated sensor data and fuzzy neural network real-time prediction model is proposed. The technical applications include an adaptive and proactive seat belt system, adaptive side airbag system and adaptive occupant neck protection system based on real-time prediction of occupant injury. The vehicle and occupant status are perceived by the system in real time, then occupant injury risk is assessed based on complex multivariate data, to solve the existing restraint system cannot adapt to different kinds of occupants and complex working condition under active and passive fusion. As a kind of integration of intelligence, networking and safety, the adaptive and proactive safety system has significant potential in improving the safety performance of vehicles and reducing the injury of occupants in accidents, and being well prepared for the more stringent requirements of advanced safety systems in E-NCAP and J-NCAP from 2026 and beyond, which ensures that the products remain at the forefront within an ever-evolving assessment framework.