Abstract: ［Objective］ With the development of the rail transit industry and advancements in sensing and positioning technologies, numerous low-cost data collection devices are now connected to the internet, which enables access to abundant operational data but also poses challenges in data management due to information overload. Efficient real-time data processing capabilities can provide actionable insights for decision-making systems and support various artificial intelligence algorithms used in rail transit sector. Addressing this issue, the real-time data sharing platform technology for rail transit management systems is investigated. ［Method］ The role of sharing platforms in the rail transit industry is analyzed, emphasizing their importance in improving operational efficiency and safety. The primary requirements for real-time data sharing from business systems are outlined and the technical selection criteria are derived from these requirements. Finally, a qualitative and quantitative comparison is conducted to evaluate the suitability of different middleware options, identifying the most appropriate middleware for platform implementation. ［Result & Conclusion］ Experimental results indicate that Hazelcast, a middleware based on an in-memory data grid, performs excellently well in handling large volumes of concurrent requests and high data loads, with stable response time growth. This makes it a suitable technical foundation of real-time data sharing platform for rail transit management systems. Additionally, Hazelcast demonstrates high performance, ease of configuration, and robust cluster support capabilities, effectively meeting the real-time data processing requirements of rail transit management systems.