Background: Amid the dual contexts of rural revitalization and the widespread integration of artificial intelligence, participatory rural planning faces significant challenges, including multi-stakeholder negotiation and the preservation of cultural heritage.Objective: This study aims to visualize the goals of each stakeholder, identify and address their demands, and thereby enhance the cultural and innovative qualities of landscape projects, establishing an efficient model for rural revitalization.Method: We propose the PLDG model, which leverages large language models to simulate multi-character dialogues. Through multimodal semantic extraction and iterative consensus generation, the framework produces optimized landscape plans.Results: Using Ma Jia Di Village as a case study, the PLDG model significantly improved cultural compatibility by 145.4%, increased the villagers’ participation index (PI) from 36 to 90, and enhanced the level of design innovation. Conclusion: This framework provides a novel paradigm and practical tools for human-machine collaborative rural planning, supportingmore effective and culturally sensitive landscape design.