Understanding greenhouse gases (GHG) emissions is becoming increasingly important with the climate change. Most previous studies have focused on the assessment of soil organic carbon (SOC) sequestration potential and GHG emissions from agriculture. However, specific experiments assessing tillage impacts on GHG emission from double-cropped paddy fields in Southern China are relatively scarce. Therefore, the objective of this study was to assess the effects of tillage systems on methane (CH4) and nitrous oxide (N2O) emission in a double rice (Oryza sativa L.) cropping system. The experiment was established in 2005 in Hunan Province, China. Three tillage treatments were laid out in a randomized complete block design: conventional tillage (CT), rotary tillage (RT) and no-till (NT). Fluxes of CH4 from different tillage treatments followed a similar trend during the two years, with a single peak emission for the early rice season and a double peak emission for the late rice season. Compared with other treatments, NT significantly reduced CH4 emission among the rice growing seasons (P<0.05). However, much higher variations in N2O emission were observed across the rice growing seasons due to the vulnerability of N2O to external influences. The amount of CH4 emission in paddy fields was much higher relative to N2O emission. Conversion of CT to NT significantly reduced the cumulative CH4 emission for both rice seasons compared with other treatments (P<0.05). The mean value of global warming potentials (GWPs) of CH4 and N2O emissions over 100 years was in the order of NT