Abstract: Quantitative evaluation of the service capacity of individual units within a continuously operating reference station (CORS) network is essential for the selection and optimization of reference station locations. Considering the characteristics of strip-shaped mountain road areas, such as significant elevation differences and complex path layouts, this article proposes a network structure factor model that incorporates elevation differences between reference stations to analyze the impact of network structure on positioning performance. Using the expansion project of the G5 Beijing–Kunming Expressway in Sichuan Province as a case study, a provincial CORS network, a mountainous highway CORS network, and a fused CORS network were constructed. Their structure factors, initialization time, positioning accuracy, and stability were compared to analyze the impact of different network configurations on real time kinematic (RTK) positioning performance. Experimental show that both the engineering CORS network and the integrated CORS network outperform the provincial CORS network in terms of structure factor, initialization time, positioning accuracy, and stability. Among them, the integrated CORS network demonstrates the highest accuracy and consistency. The findings indicate that a well-planned CORS station layout can significantly enhance the performance of RTK positioning services in CORS networks.