The rapid urbanization has resulted in the huge loss of agriculture land and water bodies, and thus has led to the deterioration of urban environment. Remotely sensed imagery has long been utilized to monitor spatial and temporal urban growth. This paper presents a methodology to monitor the shrinking 27 lakes in Wuhan from 1986 to 2002 by using remote sensing and GIS, based on various resolution imageries ranging from 10/20m SPOT PAN/XS to 5/10m SPOT5. Firstly, SPOT5 imageries were geo-referenced according to Gauss-Kruger projection with Krasovsky spheroid, by using 1:2000 topographical maps. Secondly, the rectified images (SPOT5) were used to geo-reference the rest (1986, 1995, and 2000 SPOT PAN/XS). The image processing is implemented through Erdas Imaging 8.7 package. The RMS error is controlled under the limit of 0.4 pixels. Thirdly, the spatial and temporal boundaries of 27 lakes locating in urban area are extracted by using normalized difference water index (NDWI), and their areas are calculated and can be compared for further analysis in a later stage in ArcGIS. This study has shown that these lakes are shrinking rapidly as stimulated by urban growth, and many of them have been drained and transformed into residential use. The results from further exploratory spatial data analysis indicate that three proximity variables (distance to urban centers, distance to major roads, and distance to water bodies) are the major spatial determinants in the spatial and temporal changes of these lakes.