China is the largest country in hydroelectric capacity to generate electricity in the world, and hydropower contributed greatly to the booming economy in the past years, which would play a crucial role in the energy-saving and emission-reducing plan. However, the traditional standpoint that hydropower was 'green energy' was caused doubts with the development of the research on greenhouse gases (GHGs) emissions from man-made hydropower reservoirs. The Three Gorges Reservoir (TGR), currently the largest hydroelectric reservoir in the world, has drawn much attention by researchers on its GHGs emissions. In this study, methane (CH4) and carbon dioxide (CO2) samples were taken in the water surface and the drawdown area using the static chambers at Yunyang, Wushan, Zigui, and Sandouping for 15 months (November 2009 to January 2011). Environmental factors, including climatic parameters, hydrological factors, water quality, and soil physical-chemical characters, were examined simultaneously with CH4 and CO2 flux observation. The results were as follows. (1) The average annual CO2 flux was 163.3 ± 117.4 mg CO2 m-2 h-1 at the reservoir surface, which was larger than the CO2 flux in most boreal and temperate reservoirs but lower than those in tropical reservoirs. Similar to the seasonal variation in CH4 flux, the CO2 flux started to increase in January and went on rising until peaking in the summer (June-August); CO2 flux gradually decreased thereafter. The seasonal variation in CO2 flux was influenced by temperature, water velocity, and pH. However, the spatial variation in CO2 flux was different from that of CH4 flux, with the largest flux measured at Wushan (221.9 ± 130.3 mg CO2 m-2 h-1), in the middle of the TGR, and the smallest flux measured at Zigui (88.6 ± 52.5 mg CO2 m-2 h-1), near to the TGD; these differences were related to the average water velocities at these different sites. (2) The average annual CH4 flux was 0.62 mg CH4 m-2 h-1, including 0.27 ± 0.15 mg CH4 m-2 h-1 of diffusion and 0.35 ± 0.57 mg CH4 m-2 h-1 of ebullition, which was similar to CH4 emission from temperate and boreal reservoirs but significantly lower than those from tropical reservoirs. Seasonal variation in CH4 flux showed that CH4 flux reached the maximum in summer and turned to the low levels in the other seasons, which was attributed to the seasonal variations in water velocity, temperature, and dissolved oxygen. Moreover, the yearly average CH4 flux decreased from the upstream to the downstream before the Three Gorges Dam (TGD), but CH4 emission from the surface of downstream river was higher than that from the surface at Zigui, the upstream water before the TGD, which might be related to the water velocity, deposited sediment, and disturbance by turbines. (3) The drawdown areas of fallow lands, croplands, and deforested lands were CO2 emission sources (108.37 ± 98.80 mg CO2 m-2 h-1). When the drylands were exposed in the drained season, the average CO2 emission from soil respiration (159.25 ± 142.37 mg CO2 m-2 h-1) was higher than that at the water-air interface in the inundated season (89.08 ± 66.84 mg CO2 m-2 h-1). CO2 flux was influenced by air temperature, soil temperature at 5 cm depth, soil moisture, and the concentration of NO3--N at the surface soil during the drained season, while was related to TIC, TC, pH, and DO during the inundated season. (4) The annual average CH4 flux was 0.23 mg CH4 m-2 h-1 (0.16 mg CH4 m-2 h-1 of diffusive flux plus 0.068 mg CH4 m-2 h-1 of bubble flux) in the drawdown area of TGR in 2010. Except rice paddies, the drawdown areas studied here were sources in the inundated season (0.22 ± 0.26 mg CH4 m-2 h-1) and a sink in the drained season (-0.008 ± 0.035 mg CH4 m-2 h-1). The average CH4 flux across the entire sampled season was 2.61 ± 3.76 mg CH4 m-2 h-1 in the rice paddy, which was the highest among the four types of land that were examined. For the three types of drylands in the drawdown area, the annual average CH4 emission from the croplands (0.12 ± 0.22 mg CH4 m-2 h-1) was significantly lower than that of the fallows lands (0.19 ± 0.28 mg CH4 m-2 h-1) and the deforested lands (0.18 ± 0.24 mg CH4 m-2 h-1), which were positively related with the duration of the inundated season. (5) The total GHGs emission ranged from 0.5794-1.168 Tg C-CO2 eq. yr-1 in 2010, which was only 2.4-4.9% of GHG would release by thermo-power plants if the same amount of electricity (84.7 billion kWh) was generated. Carbon budget at TGR showed that the carbon input to the TGR was estimated to be 10.78-13.66 Tg C yr-1, which was larger than the carbon outflow (10.65-11.24 Tg C yr-1).