Abstract: To address the issues of cumbersome tools, low efficiency of manual inspection, high risk of violations, and insufficient supervision methods in traditional railway maintenance operations, this paper designs a railway maintenance operation assistance system. By integrating Beidou high-precision positioning with intelligent track gauge technology, the system achieves centimeter-level positioning binding with track data, enabling precise identification and visualization of defects. Utilizing You Only Look Once (YOLO) algorithm-based visual recognition technology, it automatically verifies personnel identity, supervises operational behavior, and issues real-time alerts for anomalies. It constructs an intelligent risk assessment model to quantitatively evaluate risks and implement hierarchical control. Through full-process standardized operation control, it regulates workflows and clarifies responsibilities. Combined with multi-source data fusion visualization technology, it integrates operational data and dynamically presents safety situations. Pilot applications demonstrate that the system improves efficiency during maintenance windows, reduces violation rates, establishes a full-process intelligent operation management model, and provides a reference for the intelligent transformation of safety production in other railway specialties.