The analysis of urban land-cover classes and their spatial patterns are important problems in urban ecology, especially in ecologically fragile coastal areas. It is of great significance to those who optimize urban functional zones and are involved in urban planning management and sustainable development. High-resolution imagery has become an instrumental data source for detailed urban spatial pattern analysis, due to the complex structure of urban land covers made it difficult to achieve accurate and detailed information for urban landscape using medium or low-resolution images. In this study, based on China’s GaoFen-1 (GF-1) high spatial resolution remote sensing images and a reference dataset, an information extraction technology based on a combination of pixel-, object-, and knowledge- based methods is developed to classify the land covers in urban built-up areas (BUAs) of Shanghai, China. The mapping and landscape pattern analysis of urban BUAs in Shanghai has been completed based on the results of land cover classification. The experimental results show that the overall accuracy and Kappa coefficients of the land-cover classification in Shanghai urban BUAs are 83.7% and 0.71, respectively, which provide effective and reliable data for spatial mapping and landscape pattern analysis. Through the landscape analysis of the classification results of land cover in Shanghai, the results demonstrated that not only is there a high degree of exploitation and utilization of land resources in Shanghai but also that the spatial distribution pattern of land cover types is reasonable, which indicates that its development is sustainable.