This paper first introduces the improved Dijkstra’s algorithm and graph theory to analyze the effect of network structure construction and the effect of shortest path calculation of highways in toll road network of Guangxi Province; after that, using Data Envelopment Analysis (DEA), a model of toll road operation efficiency is constructed to analyze the trend of change and the characteristics of spatial distribution of the operation efficiency of toll roads in 29 provinces of China from 2020 to 2024. After that, we constructed a toll road operational efficiency model using the data envelopment analysis (DEA), and analyzed the trend and spatial distribution characteristics of toll road operational efficiency from 2020 to 2024 in 29 provinces in China. The results show that: the inter-city and county connectivity of the toll road network in Guangxi is better (the average degree is 14.5014), the inner cities and counties are closely connected (the clustering coefficient is 0.7931), and the proximity of each node of the highway network is larger (0.3191~0.6387), and the nodes have a better connectivity efficiency, especially Nanning City has the smallest average shortest distance between Nanning City and other nodes. When the Euclidean distance between the starting site and the destination site is small, the distribution of the shortest path ratio is more dispersed; with the increase of the Euclidean distance, the distribution of the shortest path ratio is gradually centralized, and the value is taken on both sides of 1.228. There is no obvious trend of growth in the operational efficiency of the toll road in the period of 2020-2024, and it shows a distribution pattern of gradual decrease from east to west. The operational efficiency of toll roads has a strong positive spatial correlation, and the eastern and northwestern regions show significant “high-high” and “low-low” clustering characteristics, respectively.
