In order to improve the city network mining method, the inter-cities’ connection strength, structure and density, and distribution patterns of city network in the Yangtze River Economic Belt of China have been empirically analyzed through the combined application of SNA method, “Dual-direction time distance” modified gravity model and ArcGIS geographic visualization method. The results show that the modified gravity model can better reveal the interaction differences between cities and reflect the current and potential economic, population and resource relations among cities. The city network density of this area has positively close relationship with the regional economic development level. The average value of degree centrality in the basin is high, but the difference between cities is obvious. The “agglomeration effect” of the central cities is significant, and the urban connections have an obvious cluster structure, showing an “M” shaped spatial distribution along the Yangtze River; The inner interaction strength of city network subgroups is high, but the connection between subgroups is low. The law of “downstream > midstream > upstream” also appears on the closeness centrality and betweenness centrality. In the future, it is essential to improve the integration and multi-level connections of urban agglomeration in the river basin and form a development pattern of “downstream driving - midstream transition - upstream connection”; strengthen the functions and connections of central and subcentral cities.