Urban land-use pattern occurs not only at certain spatial scale but also at certain classification system. Study of spatial scale effect on urban land-use pattern analysis for construction of spatial model of urban land-use changes entails consideration of urban land-use classification system selected. The objective of this paper is to investigate the effect characteristics of spatial scale on the result of urban land-use pattern analysis for two different classification systems in terms of spatial autocorrelation and fuzzy mathematics by an empirical study in the CBD (Central Business District) of Tokyo so as to provide some useful information for construction of multi-scale or hierarchical spatial model of urban dynamics. The results show that while spatial autocorrelation of all the categories of urban land-use at different classification systems in this study is scale-dependent, characteristics of the spatial scale effect on pattern of urban land-use categories are different for different classification systems. However, the general change trends of spatial autocorrelation of urban land-use for different classification systems show similar across a range of scale. Reducing number of urban land-use categories for one classification system may diminish the loss of information of land-use area across the range of scale and the effect of spatial scale on urban land-use pattern analysis in the certain extent.