A green infrastructure assessment (GIA)-constrained cellular automata (CCA) model was developed to integrate habitat preservation with spatial-temporal simulation and the corresponding establishment of an urban growth boundary (UGB). A case study of the Hangzhou Metropolitan Area successfully demonstrated the application of the GIA-CCA model. Multitemporal land use maps interpreted from remotely sensed images taken in 2000, 2005, 2010 and 2015 were utilized for validation and UGB prediction for 2020. The GI scenario incorporating the ecological network and a baseline scenario (BS) involving a current growth trend of no environmental constraint were designed to further investigate the unique advantages of the GIA-CCA model over the CCA model. Comparative analysis with groups of landscape metrics indicated that the GI scenario improved the connectivity among ecological elements, reducing urban fragmentation and promoting growth compactness, yet many irregular forms of edge sprawl and main city expansion were observed under the BS. Based on the GI scenario, the final UGBs were established, covering the central city, six new cities and nine towns. In contrast to the proposed UGBs, the planned UGBs generated by economic policies might play a negative role in preserving natural resources and responding to development uncertainties and could be improved by considering the forecasted results based on the GIA-CCA model as a helpful decision-making tool. The model could become an increasingly important part of growth management, particularly for developing countries and regions that are exposed to environmental deterioration due to rapid urbanization.