Soil respiration (R_s) from urban ecosystem plays a significant role in local and national carbon budget and cycling. Soil respiration under weather break as an important component of annual carbon balance, may have different mechanism responding to environment factors compared with that under normal weather condition. Therefore, quantifying soil respiration under weather break and determining its key environment factors are crucial for both understanding the belowground carbon balance and building carbon cycling model. However, little information about soil respiration under weather break among different ecosystems is available. Specifically, the data about soil respiration from urban ecosystem is scarce, although urban ecosystem had become a major settlement of human and previous studies have highlighted that urban ecosystem exhibited a high rate of soil respiration. In this study, soil respiration and it components in response to a weather break, cold-air outbreak, were investigated by LI-COR-8100-103 system in two typical urban green spaces, a woodland (Araucaria heterophylla) and a turfgrass (Zoysia matrell), in the Minjiang Riverside Park of Fuzhou City (26°03' N, 119°15' E), Fujian Province, China, during the middle February in 2011. The two components of soil respiration, microbial (R_h) and root respiration (R_r), were partitioned by trench method. In both ecosystems, cold-air outbreak significantly decreased soil temperature, and inhibited R_s and its components. The degree of decrease of soil temperature in woodland and turfgrass were 7.4℃ and 5.5℃, respectively. And soil respiration of woodland and turfgrass dropped by 79.4% and 71.2%, respectively. However, with the accompanying precipitation, these soil CO₂ fluxes also showed great fluctuation. In woodland, the daily mean value of R_s, R_h and R_r were 2.54 μmol · m^-2 · s^-1, 0.76 μmol · m^-2 · s^-1 and 1.78 μmol · m^-2 · s^-1, respectively, while those from turfgrass was 1.07 μmol · m^-2 · s^-1, 0.83 μmol · m^-2 · s^-1 and 0.24 μmol · m^-2 · s^-1, respectively. The variations of these fluxes were driven by soil temperature and exhibited an exponential relationship with it. The changes of soil CO₂ fluxes in turfgrass were more sensitive to soil temperature, due to the relatively simple canopy structure and lower chilling-tolerance. The temperature sensitivity, namely the Q₁₀ value, of R_s, R_h and R_r of turfgrass was all stimulated by cold-air outbreak, up to 4.18, 8.17 and 18.17, respectively. On the other hand, these soil CO₂ fluxes were also influenced by soil moisture and precipitation. In turfgrass, R_s, R_h and R_r all showed a negative linear relationship with soil moisture or precipitation (R²≥0.45, P<0.05), while there was a quadratic function relationship between R_h and soil moisture in woodland. This difference may result from the relatively high soil moisture in turfgrass due to irrigation management before measurement. However, any factor alone could not predict these fluxes very well. Multiple regression equation containing soil temperature, soil moisture and precipitation was, however, more predictive for these fluxes than any single-factor equation, suggesting that these soil CO₂ fluxes were simultaneously affected these combined factors during this period. Our this work was a preparatory study of winter soil repartition research on urban ecosystem. For a better understanding of belowground carbon cycling and a comprehensive assessment of soil carbon budget of urban ecosystem, further research is needed of simultaneous measurement of gross canopy CO₂ uptake rate, as well as the different response of soil CO₂ fluxes to environment factors between weather break and normal weather should be taken into account in building carbon cycling model of urban ecosystems.