Rapid urbanization has been recognized as a critical process in urban areas. This research focuses on the comparison of landscape dynamics in two petroleum-based cities: Houston, TX, and Daring, Heilongjiang Province, China. Based on the multitemporal satellite images, a general trend of landscape change was revealed in these two cities: natural landscapes such as grassland and wetland were degraded into a more heterogeneous pattern, while the human landscapes such as residential area expanded and replaced other natural classes gradually. To further investigate the spatio-temporal process in these two petroleum-oriented cities, the Markov stochastic model was adopted to simulate their landscape dynamics. The results of this study provide evidence of the threatened natural landscapes and environment deterioration during the urbanization processes in two petroleum-oriented cities. This research also demonstrates that the integration of remote sensing and GIS model is an effective approach for comparing spatial pattern and temporal process of urban landscapes.