Abstract: Against the backdrop of pacing up economic globalization and regional economic integration,the international elements flow more freely and frequently,such as flow of labor,material,funds,technology and information.The tends promote the development of regional society and economy along with the evolution of urban spatial structure,and make the spatial combination among urban agglomerations act more alive with a character of net connection.As the suburbanization develops,spatial diffusion gradually turns into a new study field.The spatial association of urban agglomerations is an abstract concept,which we defined it not only a connection among cities of urban agglomerations but also a connection among urban agglomerations as a whole and outside regions.By applying the economic relation intensity model,urban flow model and urban accessibility model,this article constructes a series of spatial combination capacity model and analyzes the spatial differentiation characteristics of spatial combination capacity taking the mid-southern Liaoning Urban agglomerations as an example.Based on the analysis,the SOM neural network grading model is built to evaluate the spatial combination capacity of the ten node cities.The results shows that: 1) Shenyang has the largest total economic linkage.The relationship between economic relation intensity of Shenyang and other cities,and the distance of railway among them presents the "S" curve.2) According to the value of urban flow intensity,the ten node cities are divided into three sorts—high,middle and low.Dalian′s value is the highest and Yingkou shows its potential as the center in the middle part of city group.3) The accessibility of the top three cities is high along the Shenyang-Dalian Highway.4) The classification result of SOM neural network indicates that Shenyang has the strongest spatial combination capacity and it shows the centrality as a solely class.