Accelerating urbanization has made noise pollution a critical constraint on the high-quality development of urban spatial quality, necessitating eff ective planning and management. Existing physical noise prediction models struggle to capture the complex, nonlinear relationships between urban morphology and noise distribution, and there is a lack of design tools for noise attenuation. To address this, this study develops a high-precision noise prediction model and proposes an urban morphology parameter- driven analysis and design paradigm aimed at “governing noise with form”. Using the central urban area of Lyon, France, as the study area, a high-precision spatial database containing 27 morphological parameters and noise distribution data was established on a big data platform. Building upon this, a deep learning prediction model with U-Net as the core architecture was constructed. Ablation experiments were employed to quantitatively identify dominant parameters; Geographically Weighted Regression (GWR) was used to locate hotspot areas for regulation; and, fi nally, an interactive visualization platform was built using the Flask framework. The results indicate that: (1) The optimal U-Net model demonstrates robust predictive performance and high spatial fi delity on an independent test set. (2) The ablation experiments identifi ed 13 dominant parameters, revealing that roads are critical noise sources and the built environment tends to exacerbate noise levels, while parameters characterizing area size and spatial aggregation in green space morphology contribute most signifi cantly to noise reduction. (3) GWR analyzed the spatial heterogeneity of the infl uence of dominant parameters, which, combined with the Flask interactive platform, realized the real-time simulation and visualization of parameter-driven noise reduction strategies. In summary, this study develops an intelligent prediction model and an interactive design tool to provide a scientifi c basis and technical support for spatial morphology optimization aimed at the construction of tranquil cities.