The dispatching operation regulations have great significance in guiding the failure handling of hydro power stations. Therefore, the dispatching operation regulations of the hydro power station are used as the research object, and a top-down construction method of the dispatching operation knowledge graph of hydro power stations is proposed by means of knowledge representation, knowledge extraction, and structured management of regulations. First, through term, concept, and relationship extraction, the schema layer of the knowledge graph is constructed. Then, the data layer is built according to the schema layer with entity extraction of regulations by using a deep learning model with a bi-directional gated recurrent unit (BiGRU) network and conditional random field (CRF). Finally, by learning the dispatching operation regulations of a large-scale ladder hydro power station located in China, the dispatching operation knowledge graph of the hydro power station is constructed, and the effectiveness of the proposed method is verified with simulation results. The results show that the constructed dispatching operation knowledge graph of the hydro power station can help the staff of the hydro power station to carry out failure handling with assistant decision-making and effectively improve the level of emergency management and dispatching intelligence of the hydro power station.