This study proposes a general data collection system architecture based on relational database technology, aimed at meeting complex and diverse data collection needs. Based on the characteristics of relational databases, the system achieves dynamic correlation and integration of multi-source data through the “knowledge element” model, and designs a universal underlying storage structure that supports multiple collection tasks. By adopting object serialization method, the compatibility issue of different data storage modes has been solved, enabling unstructured data to be stored in a unified JSON string format in relational databases. In addition, based on the logical model of the “self association” table, multi-level forwarding support for deterministic and additive collection tasks has been implemented, and a state transition model between hierarchical tasks has been constructed. In the specific implementation, ASP NET WebForm technology combines transaction scripts and activity recording patterns to dynamically generate data collection interfaces; Implement bidirectional binding between front-end and data through Vue framework, and use Web API to complete dynamic data loading and storage. To solve the common storage problem of tree structured data, the system introduces Level and Path fields, and combines triggers to automatically maintain values, significantly improving the efficiency and automation level of complex data collection tasks.