Vernacular landscapes, regarded as intricate systems, are distinguished by their intrinsic dynamism and uncertainty. The different preferences of stakeholders frequently result in varied evolutionary trajectories, thereby presenting challenges for planners to reconcile multiple strategies and precisely evaluate the future consequences of various development pathways. To augment students’ comprehension of complex systems and their capacity to make integrated decisions within vernacular landscape education, this study introduces a planning studio case situated in the urbanizing polder landscapes of Nansha District, Guangzhou. We propose a digital design-assisted methodology grounded in Agent-Based Modeling (ABM) to simulate and assess various prospective development scenarios. The model incorporates the degree of implementation of three strategic approaches, natural development, zoning planning, and traditional landscape pattern preservation, as key variables, generating 25 distinct evolutionary scenarios. Comparative analysis of landscape multifunctionality across these scenarios reveals that strategies dominated by traditional landscape patterns achieve the highest level of sustainability. Building upon the modeling results, student teams further completed detailed spatial design proposals. This study demonstrates the potential of ABM in addressing complexity in vernacular landscape design, enhances the capacity of digital tools to support planning decisions, and offers valuable insights for both landscape planning education and professional practice.