An integrated model for spatial association metrics is developed on the basis of fractal correlation coefficient, global and local spatial association measurements, and semivarigram. And six input data for the model are acquired by multi-source satellite imageries covering the period 1984-2005, including Landsat TM or ETM, IRS-P6. Then, we used the model to study the urban growth periodicity, the fractal dimensions of the built-up area, the spatial distribution and evolvement of the "hot spot" districts of urban growth as well as the pattern evolution of spatial gradient since the reform and opening-up of China. Finally, some conclusions are drawn as follows. (1) Since the 1980s, Suzhou's urban growth has experienced four stages from a gradual increase, a slow decrease, then a rapid rise to a sharp decline. (2) There exists an obvious characteristic of spatial evolution and shift according to the analysis of high-value clusters, viz. "hot spots", of urban growth intensity. The hot spot of urbanization is in turn from 1984 to 2005, going from Suzhou city, Suzhou city-Wujiang axis to Suzhou city-Kunshan axis along the Shanghai-Nanjing expressway, and then to Taicang-Changshu-Zhangjiagang axis along the Yangtze River. These hot spot regions are the cores driving social and economic development in differenrt periods in Suzhou. In the development period, spatial continuity and self-organization of urban growth has been intensified, and the spatial form of urbanization changed from decentralization in the initial stage to centralization at present. Additionally, the structural difference due to spatial association becomes more and more obvious with a characteristic of strip association, showing a tendency of a regional urbanization. It is worth noting the fact that the driving force of urbanization is different in different stages.