Spatial data analysis is the key component of GIS. Current GIS systems focus on analytical functions for geometric operations, but few spatial statistical functions were used in GIS. This paper examined the feasibility of combining GIS and spatial statistical techniques in the analysis of land use changes in the coastal area of Yunlin County, Taiwan. Markov model was employed to analyze land use changes for three 5-7-year time intervals during 1976-1992. The spatial pattern analysis went a step further by cluster analysis to derive the change pattern over space and time. The result indicates different patterns and trends of land use changes among three different land use types. This result cannot be derived solely by either GIS functions or statistical analyses. It can only be derived by integrating GIS and spatial analysis. The applicability of integrating spatial statistics and GIS has been explored.