The occurrence of railway operation accidents is often caused by a series of hazards. It is necessary to predict hazards before accidents so as to prevent railway operational accidents. In this paper, a new method of knowledge graph embedding-based railway operational hazard prediction is proposed. Specifically, a ConvKB-based hazard prediction approach is used for the specific link prediction. Its innovation is applying the knowledge graph embedding-based method to railway operational hazard prediction. The case based on railway operational accidents in the UK is studied, and the results show that this method can effectively predict the railway operational hazards. The method proposed is helpful to formulate some targeted control strategies and thus achieve the purpose of preventing railway operational accidents.