Based on a large number of texts produced by railway traction power supply equipment during service, which is difficult to analyze and apply directly, a text information mining system framework for railway traction power supply equipment operation and maintenance is proposed. This paper covers the process of Data pre-processing, information extraction, analysis and application, and studies the key technologies of Chinese text segmentation, information extraction, status evaluation, etc. Finally, the system architecture and related key technologies are verified by actual traction power supply equipment operation and maintenance text data.