Urban lawns have expanded rapidly with urbanization, substantially influencing regional carbon budgets. However, research on their greenhouse gas (GHG) source-sink dynamics remains scarce and the driving mechanisms are poorly understood. This study investigated CO₂, CH₄, and N₂O fluxes from urban lawns beneath typical broad-leaved (BF) and coniferous (CF) forests in a major city in the Yangtze River Delta (YRD) region, China. The seasonal and diurnal patterns of GHG fluxes were characterized, and the roles of environmental factors and mowing management were analyzed. Results showed that CO₂ fluxes from both lawn types exhibited pronounced seasonal fluctuations, with higher emissions in summer and lower in winter. The mean annual CO₂ flux from CF lawns (17.805 ± 1.016 mg m⁻² min⁻¹) was significantly higher than that from BF lawns (7.643 ± 0.881 mg m⁻² min⁻¹). Diurnally, CO₂ fluxes displayed a "daytime low, nighttime high" pattern, contrasting with the typical "daytime high, nighttime low" unimodal pattern commonly observed in grassland soils. Annual mean CH₄ fluxes for the BF and CF lawns were −0.452 ± 0.027 μg m⁻² min⁻¹ and −0.565 ± 0.035 μg m⁻² min⁻¹, respectively, indicating net CH₄ uptake throughout the year, with soil volumetric water content (VWC) emerging as the key common regulatory factor. Diurnal CH₄ variations followed a "daytime high, nighttime low" pattern in spring and summer but showed no clear diurnal rhythm in autumn and winter. Annual mean N₂O fluxes were 0.212 ± 0.020 μg m⁻² min⁻¹ and 0.172 ± 0.022 μg m⁻² min⁻¹ for the BF and CF lawns, respectively, with higher emissions in spring-summer and lower in autumn-winter, indicating net N₂O emission overall. N₂O also deviated from the typical diurnal pattern, likely due to pronounced fluctuations in soil water content. Mowing management showed that spring mowing caused the least disturbance to GHG fluxes, summer mowing significantly enhanced CO₂ emissions, and autumn mowing promoted CH₄ uptake and N₂O emission. By combining structural equation modeling (SEM), random forest (RF) importance ranking, and SHAP analysis, we identified fundamentally distinct environmental driver regimes between the two canopy types: the BF lawn exhibited a "temperature–inorganic nitrogen dominated" pattern, whereas the CF lawn showed a "moisture–microbial–carbon pool co-regulation" pattern. Overall, urban lawns exhibit a unique GHG source–sink pattern and regulatory regimes distinct from those of natural grasslands. Quantifying their GHG budgets is of considerable scientific value for optimizing urban carbon management and advancing low‑carbon city development.