At present, smart contracts are designed based on the application field, and their structure and functions are closely related to specific businesses. Even smart contracts within the same field and the same business have different structures and functions due to different developers, resulting in structural confusion, repeated development, and low levels of sharing. In response to this problem, this study conducts a full investigation of smart contracts in various fields, using big data technology to compare and analyze the structures of each contract and extracting the common content of their main bodies to study each feature, as well as to conduct induction and fusion. This study also generally designs a hierarchical structure and formulates structural modules such as transaction rules and the analysis of rights and responsibilities, as well as a reward and punishment mechanism. Data traceability is established, and the overall architectural specification of smart contracts is constructed. Additions, deletions, and improvements are made based on specific application environments to realize the dynamic updates of the architecture of contracts. Experiments show that the architecture of contracts can realize the various functions required in a transaction, solve the problem of the repeated development of current transactional smart contracts, and improve the sharing level.