Artificial intelligence (AI) technology is increasingly applied in healthcare, showcasing unique advantages in processing complex perioperative clinical data characterized by multi-source heterogeneity and temporal dynamics. This study, leveraging clinical practices at Guangdong Provincial People’s Hospital’s Anesthesiology Department and HiFly’s medical IoT technologies, implements the Technical Specification for Perioperative Data Acquisition and Governance to enable standardized perioperative IoT data collection. A patient-centered interdisciplinary clinical database (WIRE) was constructed by integrating time-series and relational databases, supporting multimodal data access, standardized management, and highly efficient queries. An AI integration engine was developed to facilitate quick model deployment and secure access for clinical decision support. The system has achieved comprehensive perioperative monitoring, early warning, and knowledge reasoning, accumulating 1.6 billion high-quality data records to date. It has established crucial AI applications, including early warning models for “three lows and one spasm” during surgery (hypotension, hypothermia, hypoxia, and convulsions), significantly promoting intraoperative risk identification and intelligent intervention. This system offers a pioneering paradigm for smart anesthesia and refined perioperative management.