National key towns are the leader in the development of small towns. The formation of scientific and reasonable national key town layouts is of great significance for optimizing China's urbanization strategy. Taking the 1887 and 3675 national key towns published in 2004 and 2014 respectively as samples, this study used geodector and the geographically weighted regression (GWR) model to explore the changing characteristics of their distribution and analyze the influencing factors of national key town layouts and effects at the prefecture level. The results show that: 1) After the adjustment by the government, the balance of the layout and construction effect of national key towns is enhanced. The main gathering areas moved westward and northward, the distribution of cold spots and hotspots break through the "Hu Huanyong Line", and the degree of differentiation of economic radiation effect is weakened, reflecting the powerful influence of policy factors. In addition to the inter-county balance and regional preferential policies, the distribution of national key towns is also significantly affected by factors such as altitude, road network density, and resident population urbanization rate. 2) The combination of the factor detector, the GWR model, and the interaction detector can more accurately characterize the mode of action, direction, path, and intensity of the influencing factors. The layout of national key towns is not the result of the uniform, independent, and direct effect of the five significant factors, but the product of the synergistic effect of the interaction of the factors with spatial heterogeneity. 3) The synergy between the inter-county balance policy and other factors is the dominant force in the formation of the existing layout of the key towns; the effect of the regional preferential policy is overall positive, but the target areas need to be more accurately identified.

Keywords： national key towns ; quantitative attribution ; Geodetector ; geographically weighted regression ; China