Urban lakes are hotspots of methane (CH4) emissions. Yet, actual field measurements of CH4 in these lakes are rather limited and our understanding of CH4 response to urban lake eutrophication is still incomplete. In this study, we measured dissolved CH4 concentrations and quantified CH4 diffusion from four urban lakes in subtropical China during wet and dry seasons. We found that these lakes were constantly CH4-saturated, contributing the greenhouse gas (GHG) to the atmosphere. Nutrient enrichment significantly increased CH4 concentrations and diffusive fluxes. Average CH4 flux rate in the highly-eutrophic lake zones (4.18 ± 7.68 mmol m−2 d−1) was significantly higher than those in the mesotrophic (0.19 ± 0.18 mmol m−2 d−1) and lightly/moderately-eutrophic zones (0.72 ± 2.22 mmol m−2 d−1). Seasonally, CH4 concentrations and fluxes were significantly higher in the wet season than in the dry season in the mesotrophic and the lightly/moderately-eutrophic lake zones, but an inverse pattern existed in the highly-eutrophic lake zones. CH4 concentrations and fluxes increased with elevated levels of nitrogen, phosphorus and dissolved organic carbon (DOC). The accumulation of nutrients provided autochthonous substrate for methanogenesis, indicated by a negative correlation between CH4 and the C:N ratio. Ammonium-nitrogen (NH4+-N) was the best predictor for spatial fluctuation of CH4 concentrations and diffusive fluxes in the mesotrophic and the lightly/moderately-eutrophic lake zones, while total nitrogen (TN) and total phosphorus (TP) levels showed the highest predictability in the highly-eutrophic lake zones. Based on the findings, we conclude that nutrient enrichment in urban lakes can largely increase CH4 diffusion, and that urban sewage inflow is a key concern for eutrophication boosting CH4 production and diffusive emission. Furthermore, our study reveals that small urban lakes may be an important missing source of GHG emissions in the global C accounting, and that the ratio of littoral-to-pelagic zones can be important for predicting lake-scale estimation of CH4 emission.