Abstract: To explore methods for identifying and classifying the urban spatial structure of mountainous counties, and to facilitate spatial structure optimization and the formulation of urban development strategies for county-level regions, this study selects Southern Shaanxi as the research area. It integrates the gravity model with social network analysis to portray the spatial network of county urban systems. Evaluation indicators are selected from two dimensions, namely regional agglomeration capability and intra-county connection network, to establish a methodological system for identifying spatial structural characteristics. Adopting the K-means clustering method, this paper divides the urban spatial structure of counties in Southern Shaanxi into five types: central hub type, subcentral hub type, general development type, sub-peripheral polar-core type and peripheral polar-core type. Targeted spatial optimization strategies are proposed for each type, which can provide a theoretical basis and practical references for the optimization of county urban system and the high-quality development of towns in mountainous areas of Southern Shaanxi.