Intelligent connected vehicles, as the focus of the global automotive industry, are currently at a critical stage of large-scale commercialization. However, during the development process of vehicles from mechanical systems with limited functions to mobile intelligence with complex and multiple functions, the issues of functional safety, cybersecurity, and safety of the intended functionality are the main challenges of the industrialization of intelligent connected vehicles, including multiple safety risks such as hardware and software failures, insufficient performance in edge scenarios, cyber-attacks and data leakage. In this paper, the safety and security issues of intelligent connected vehicles, the challenges posed by emerging technology applications, and related solutions are systematically reviewed and summarized. A fusion safety system framework with the safety cube as the core of protection and control is proposed innovatively based on a field-vehicle-human safety interactional model, realizing stereoscopic, deep, and comprehensive safety protection through end-cloud collaboration. Meanwhile, an X-shaped fusion safety development process based on CHAIN is proposed. Through the empowerment of digital twin and AI technologies, it could approach interaction between physical entities and digital twin models and the automation of the development process, thereby satisfying the demands of fusion safety system design, intelligent development, rapid delivery, and continuous iteration. The fusion safety system framework and X-shaped development process proposed in this paper can provide important insight into intelligent transportation vehicles and systems' safety and security design and development.