The urban impervious surface (IS) refers to any nonporous land cover that prevents water from infiltrating into sub-surface layers, e.g., buildings, roads, parking lots, sidewalks, and other built surfaces. In addition to its use as an indicator of environmental influences, IS has also been applied to determine the spatial extent, intensity, and type of urban land use/cover changes. In recent years there has been increased interest in the use of classification and regression tree model (CART) technology to map sub-pixel impervious surfaces. This process uses medium-resolution Landsat imagery to extrapolate IS over large-scale areas with high-resolution imagery as training data to represent the urban land-cover heterogeneity. The main advantage of the regression tree algorithm is that it can account for non-linear relations between predictive and target variables, and thus allows both continuous and discrete variables to be used as input (predictive) data. The distribution of impervious surface index (ISI) distribution was deriveded fom Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) data by comparing it with CART and multi-stepwise regression (MSR). The results demonstrated that CART provided with the correlation coefficient of 0.94 and the average error of 8.59% with consistent and acceptable accuracy, which was better than MSR. The average ISI value for the total study area was 20.80% with standard deviation of 0.29. However, in most grids (58.60%) the average ISI was less than 10%. ISI percentage values in different regions also varied dramatically ranging from 67.32% in Zone 1 to 9.32% in Zone 6. Further, the spatial distribution patterns of IS exhibited spatial gradients increasing in value from the city outskirts to the inner urban areas. Utilizing ISI a new landscape classification system was developed, composed of the following four categories: natural cover (ISI≤10%), low-density urban (10< ISI≤40), medium-density urban (4160). The results of landscape pattern analysis demonstrate that high-density urban is the dominant landscape within the 4th ring-road covering 67.41% of the surface area, while natural cover is the dominant form of land cover outside the 5th ring-road. Landscape patterns varied extremely with landscape fragmentation index and average patch area, and the average area of natural cover exhibits a U shape as you moved from the inner urban area to the outskirts. It can be concluded that ISI is able to serve as a useful indicator for landscape classification and landscape pattern analysis.