互花米草(Spartina alterniflora)入侵严重威胁滨海湿地生物多样性与生态功能，研究互花米草生长扩张过程对湿地生态和资源综合管理至关重要。本文构建了基于生物量概念的元胞自动机模型以模拟互花米草的时空分布，通过在计算中引入生物量，突破了传统模型仅能反映植被“有”或“无”状态的局限。模型采用摩尔邻域法则进行空间迭代计算，并综合考虑了淹水深度(表征水动力胁迫)、生长大周期(反映个体发育)、季节性物候变化等环境与生理约束。将这些因素作为修正系数动态调整元胞生长，以准确模拟盐沼生态系统中植被面积和密度变化的时空分布。选取江苏盐城湿地珍禽国家级自然保护区，以2000年5月的互花米草面积和密度分布为初始条件，对其15 a的生长扩张过程进行模拟。结果显示，2015年10月互花米草的密度分布模拟值与遥感解译数据之间的Pearson相关系数达到0.8247，表明模型在重现长期时空动态方面具有较高的可靠性与预测精度。此外，通过敏感性分析，界定了模型关键参数的合理取值范围，为模型的进一步应用提供了参数选择依据。本研究建立的元胞自动机模型不仅为河口湿地入侵物种的生态管理、资源保护及修复决策提供了强有力的技术支持，也为类似生态系统的研究提供了新的视角和方法。

The invasion of Spartina alterniflora (Smooth Cordgrass) in coastal wetlands has emerged as a pressing ecological challenge, significantly threatening native biodiversity and impairing critical ecosystem functions such as sediment stabilization, carbon sequestration, and habitat provision for migratory species. In response, comprehensive investigations into the spatiotemporal dynamics of this invasive species are urgently needed to inform effective ecosystem management and resource conservation strategies. This study addresses this imperative by developing an innovative biomass-based cellular automaton (CA) model specifically tailored to simulate the complex expansion patterns of S. alterniflora in salt marsh environments. Traditional CA models for vegetation dynamics often suffer from oversimplification, typically reducing vegetation presence to binary ‘on/off’ states without accounting for growth intensity or physiological responses. Our model advances this framework by incorporating biomass as a continuous variable, enabling nuanced representations of vegetation health, growth stages, and stress responses. The model employs Moore’s neighborhood rules for spatial iteration, which consider eight surrounding cells to more realistically capture seed dispersal and vegetative spread patterns. Key environmental and physiological constraints are integrated as dynamic correction factors: water depth (quantifying hydrodynamic stress), growth macrocycle (mapping individual developmental stages), and seasonal phenological changes (accounting for dormancy and active growth periods). This multi-parameter approach allows for accurate simulation of vegetation area and density distribution across heterogeneous wetland landscapes. The model was applied to the Jiangsu Yancheng Wetland National Nature Reserve, a representative coastal ecosystem, using May 2000 distribution data as initial conditions to simulate a 15-year expansion period. Validation against October 2015 remote sensing data demonstrated exceptional performance, with a Pearson correlation coefficient of 0.8247 between simulated and observed density patterns. This strong agreement confirms the model’s capacity to replicate long-term spatiotemporal dynamics with high fidelity. Furthermore, rigorous sensitivity analysis identified critical parameter thresholds for biomass accumulation rates and growth macrocycle, establishing robust operational ranges for model application in unstudied regions. The developed CA model represents a significant methodological advancement, offering valuable technical support for invasive species management in coastal wetlands. Its applications extend to informing restoration strategies, optimizing control measures, and supporting conservation planning. Beyond immediate practical utility, the model provides a novel analytical framework that can be adapted to study other invasive species or ecosystem types, thereby contributing to broader ecological research and management efforts. By bridging the gap between theoretical modeling and ecological reality, this study enhances our capacity to predict and mitigate the impacts of biological invasions on vulnerable wetland ecosystems.