At present, many integration construction problems such as the disconnection of integrated development of urban rail stations and surrounding areas, insufficient integration of stations and cities, insufficient accessibility, and low comprehensive benefits are becoming more and more prominent, and at the same time, the mechanism for supervising and implementing the assessment of urban rail stations is not yet perfect. Based on the framework of urban rail station assessment system, this paper established an integrated assessment index system for urban rail stations from eight dimensions, including land function, traffic organisation, spatial environment, commercial vitality, economic benefits, safety resilience, underground space and implementation effectiveness, and took 32 rail stations in Beijing as the objects of study, and carried out the assessment and classification of the rail stations in different radial scales by using entropy Weight-TOPSIS method. Evaluation and classification of different radiation scales of rail stations were carried out by using entropy Weight-TOPSIS method. The results show that there are spatial differences in the influencing factors of rail station integration construction in different radiation scales of rail stations, and the assessment results are divided into four categories by using K-means clustering, and classification optimisation suggestions are put forward.