Abstract: From the perspective of network and based on the gravity model and complex network theory, this paper studies the evolution of the railway logistics network structure in China from 2015 to 2021. The results show that:（ 1） The scale of railway logistics in China and the inter-provincial railway logistics connections are generally strengthened, and basically synchronous development between regions is achieved. Among them the acceleration of railway logistics development in Xinjiang is the most significant.（ 2） The regional development of railway logistics in China is unbalanced and there is a "Matthew Effect". The development speed of railway
logistics in Hainan, Xizang, Shanghai, and Qinghai is seriously lagging behind. （3） The centrality is significantly improved, and the
external radiation capabilites of central provinces in the railway logistics network, such as Shanxi, Nei Mongol, Shaanxi, and Shandong is enhanced.（ 4） The "Small World" characteristic is significant. From 2015 to 2021, the five major communities evolved into three major communities, with Shanxi-Hebei-Nei Mongol, Shandong-Henan-Shaanxi, and Guangdong-Guizhou-Sichuan as the main development axis chains. On this basis, optimization measures for the railway logistics network structure in China are put forward.

Key words: gravity model, complex network, railway logistics, Matthew Effect