Abstract: Power grid station wiring diagrams, as important data of the power system network structure, carry the key information of power equipment operation and play an important role in the application of power grid dispatch and control. A multi-stage collaborative intelligent recognition method for power grid station wiring diagrams is proposed to address the problem of dispatch and operation personnel manually drawing, which leads to tedious work and a large task volume. This method achieves machine-assisted drawing. Firstly, cutting detection and line fitting are used to enhance the generalization ability of graphic targets and graphic direction recognition is achieved through structural feature extraction and fuzzy comparison processing. Secondly, based on the principle of random pasting, the robustness of the text recognition model is enhanced, and the accuracy of graphic text matching is improved through knowledge fusion technology. Finally, based on connected domain detection and identification of connecting lines, traversal filtering is used to match the connection relationship between the entity and the main wiring. The experimental results show that the overall recognition accuracy of the method for the factory station wiring diagram is over 88%. When applied to the substation monitoring system and converter station operator control system, it greatly improves the efficiency of personnel composition modeling.