Traditional user participation in spatial design often encounters three fundamental challenges that affect design outcomes: (1) insufficient knowledge of relevant design principles, (2) lack of proficiency in professional design software, and (3) limited design options that fail to meet personalized needs. In response, this study proposes an integrated artificial intelligence (AI)-based approach to support user participation in the design process. Natural language processing (NLP) techniques are applied throughout the design workflow to capture user intentions, provide relevant design knowledge, and explore broader design possibilities. By incorporating tools such as Whisper, GPT-3.5, and Stable Diffusion, the approach is applied and validated through a case study of landscape renovation in the Qingnian East Road Community, Jinan. System testing results demonstrate the considerable potential of NLP-based design tools compared to traditional UI-based interaction. The average Conversations Per Session (CPS) between users and the intelligent agent reached 13, with over 85% of interactions deemed effective. The system effectively assists users in overcoming operational difficulties and provides knowledge-based suggestions for issues such as spatial configuration and material selection. While the text-to-image generation module partially fulfills users’ personalized design needs, the output quality is limited by the style and scale constraints of the existing Stable Diffusion model. Future work should focus on developing domain-specific models to better align with the diverse needs of personalized spatial design.