A decentralised Web3 framework for ecological landscape Design in Urban Parks Based on AI-aided Virtual Interaction. The framework links multi-source park data, a spatial ecological graph, AI co-design, and smart-contract governance so that design states, feedback records, and voting outcomes remain verifiable, with large scene assets stored off-chain. After ten iterations of the three renewal modes, its composite ecological benefits index was 0.95, with a mean habitat suitability value of 0.89 for all six parks; There was a relatively average gain rate of +56% in visitor satisfaction and +50% in carbon storage between new and old buildings. At the peak of five thousand simultaneous user access, the system’s response time per unit of operation were respectively eight-two (ms), four-hundred-eighty-nine (-81%), forty-seven-fortytwo-six(-77%). Ablation test results show: without consensus confirmation, eco-enforcement and virtual reinforcement the effectiveness decreased by 4.8-12.7 per cent. Although the form is not perfect yet; still provides an operational path to join ecological Design and participatory Governance.
