The digital management of engineering construction is one of the primary tasks of smart construction sites. In the current practice of construction site information management, there are a series of fraudulent behaviors such as engineering data being forged and modified and this seriously interferes with the safety supervision of engineering projects. In order to improve the security and traceability of multi-type engineering data, this paper presents a framework of system for construction site information management using blockchain and smart contracts. In this method, multiple independent smart contracts are developed for different data types. It ensures that the system can dynamically deploy new smart contracts without interrupting service, facilitating the expansion of scenario services. More importantly, the data analysis function is extended to the terminal, which allows the integration of different data analysis algorithms such as deep learning into information management. The design breaks the restriction that only rule-based data analysis can be performed. Meanwhile, the distributed federation architecture ensures that the data cannot be tampered with. Finally, the three most common types of data (namely text, time-series, and image) were used to verify the security and stability of the management system.