Methane emissions play a key role in global warming and climate change. The mechanism of methane emission in the paddy fields of the Taihu Lake Basin-a typical rice-wheat rotation region of China and also an area with one of the highest rates of land use change in the country-is still unclear, largely due to limited observation and the spatial heterogeneity of the environment. Eddy covariance, one of most popular approaches to observing methane flux, has many advantages, including long-term continuous observation, high frequency, and large monitoring range, without destruction of the original soil environment. We utilized the eddy covariance method to observe the methane flux in rice-wheat rotation paddy fields in the Taihu Lake Basin. The missing flux values were interpolated with the GA-BPNN model. The characteristics and influencing factors of methane flux were then analyzed, and a semi-empirical multiplicative model was built. The results showed that the rice-wheat rotation paddy fields were the main source of methane in the Taihu Lake Basin. During the observation period, the total amount of emitted methane was 28.95 g/m², and methane flux values varied in the range of 0-0.861 μmol m^-2 s^-1. The diurnal variation in methane flux during the rice season appears under two modes:irregular and unimodal. In the irregular mode, methane flux was unstable throughout the day, and in the unimodal mode, the emissions during the day were significantly higher than that at night. The emission of methane in the rice season is concentrated in the early and middle stages of rice growth, relatively weak in the later stage, relatively low in the green stage (0.102 μmol m^-2 s^-1 per day), strong in the tillering stage (0.451 μmol m^-2 s^-1 per day), and lowest in the maturity stage (0.006 μmol m^-2 s^-1 per day). The results of the analysis of various impact factors are as follows:The methane flux in the rice season increased exponential with air temperature and soil temperature at 10/20/40 cm depth (R²=0.589, 0.584, 0.521, 0.459, P<0.0001). Methane flux increased with increasing temperature. Methane flux showed a weak exponential relationship with soil moisture at 10/20 cm depth (R²=0.362, 0.372, P<0.0001). With the rise in soil moisture, methane flux increased. Methane flux has a quadratic relationship with soil moisture at a depth of 40 cm (R²=0.378, P<0.0001). When the soil moisture is less than 0.41 m³ m^-3, methane flux decreases with the increase in soil moisture, and when the soil moisture is greater than 0.41 m³ m^-3, methane flux increased with the rise in soil moisture. Methane flux showed a power relationship with soil conductivity at 10/20/40 cm depth (R²=0.309, 0.54, 0.439, P<0.0001). As the soil conductivity increases, the methane flux increases. Methane flux and friction wind speed are only significantly correlated on a half-hour scale. The methane flux estimated by the model 2 is 6.69% lower than measured flux. The model 2 well simulates the methane emission from paddy fields in the Taihu Lake Basin. The main driving factors of methane emission from paddy fields in the Taihu Lake Basin include soil temperature, soil moisture, soil conductivity, and friction wind speed.