<p>With the ongoing advancement of urbanization, urban parks and green spaces, which are vital public resources for enhancing the quality of life for urban residents, have garnered increasing attention. A thoughtful arrangement of green spaces can improve the urban ecological environment and effectively bolster the city’s capacity for sustainable development. Traditional methods for optimizing green space layouts primarily depend on manual planning and empirical analysis; however, these approaches often face inefficiencies and limitations when addressing large-scale urban areas. This study analyzed the distribution of green spaces in a representative metropolitan area in southern China, selecting 200 urban park samples collected between 2021 and 2023 that included data on green space coverage, surrounding environments, and population density. The UNet model was trained with this dataset to extract optimal green space layout patterns. Compared to conventional methods, the optimized layout scheme notably improves spatial utilization and distribution. Specifically, the optimized green space coverage rate increased by 12%, and the per capita green space area rose by 8%, with significant gains observed in densely populated urban districts. These findings provide data-driven insights to promote more sustainable and equitable green space planning in urban environments. Detailed analysis indicates that the optimized green space coverage rate has risen by 12%, and the per capita green space area has increased by 8%. The results of this study present a new approach for the scientific planning and optimization of urban park green spaces, holding considerable practical value and potential for widespread adoption.</p>