Trait-based approaches offer valuable perspectives for vegetation classification, but functional traits struggle to capture resource allocation among competing plants, showing limitations across scales. This study aimed to introduce plant architecture to enhance trait-based vegetation classification. From 2021 to 2023, 32 plots of Coastal Dwarf Forests (CDF) and Coastal Normal Forests (NCDF) along China’s eastern coast were surveyed. Their community characteristics were quantified, and classification and clustering models assessed the advantages of plant architecture in distinguishing these communities. The results indicated plant architecture traits are more critical for distinguishing different community types than leaf-based functional traits. Additionally, plant architecture traits are effective in clustering plant associations within the same community type. This is because plant architecture traits are closely linked to habitat, phylogeny, and community structure, providing a comprehensive description of vegetation, while functional traits reflect only partial habitat information related to soil nutrients. Our findings underscore the importance of plant architecture in optimizing trait-based vegetation classification and suggest that variations in the plasticity of plant architecture traits may support the classification of CDF as a distinct vegetation unit.