The Three Gorges Reservoir is a good site for the further researches on reservoir induced seismicity due to decades’ seismic monitoring. After the first water impounding in 2003, seismic activity becomes more frequent than that before water impoundment. In order to quantitatively study, the relationship between the water level fluctuation and earthquakes in TGR, we introduced statistical methods to attain the goal.First of all, we relocated the earthquakes in TGR region with double difference method and divided the earthquakes into 5 clusters with clustering analysis method. Secondly, to examine the impacts of water level fluctuation in different water filling stages on the seismic activity in the 5 clusters, a series of statistical analyses are applied. Pearson correlation results show that only the 175 m water level fluctuation has significantly positive impacts on the seismic activity in clusters Ⅰ, Ⅱ, Ⅲ and Ⅴ with correlation coefficients of 0.44, 0.38, 0.66 and 0.63. Cross-correlation analysis demonstrates that 0, 1, 0 and 0 month time delay separately for the clusters Ⅰ, Ⅱ, Ⅲ and Ⅴ exists. It illustrated the influences of the water loading and pore pressure diffusion on induced earthquakes. Cointegration tests and impulse response analysis denoted that the 175 m water level only had long term and significant effects just on the seismic events in the intersection region of the Fairy Mount Fault and Nine-brook Fault. One standard deviation shock to175 m water level increased the seismic activity in cluster Ⅴ for the first 3 months, and then the negative influence was shown. After 7 months, the negative impulse response becomes stable. The long-term effect of the 175 m water impoundment also proved the important role of pore pressure diffusion in RIS with time. The Three Gorges Reservoir is a good site for the further researches on reservoir induced seismicity due to decades’ seismic monitoring. After the first water impounding in 2003, seismic activity becomes more frequent than that before before water impoundment. In order to quantitatively study, the relationship between the water level fluctuation and earthquakes in TGR, we introduced statistical methods to attain the goal. First of all, we relocated the earthquakes in TGR region with double difference method and divided the earthquakes into 5 clusters with clustering analysis method. Secondly, to examine the impacts of water level fluctuation in different water filling stages on the seismic activity in the 5 clusters, a series of statistical analyzes are applied. Pearson correlation results show that only the 175 m water level fluctuations has significantly positive impacts on the seismic activity in clusters Ⅰ, Ⅱ, Ⅲ and Ⅴ with correlation coefficients of 0.44, 0.38, 0.66 and 0.63. Cross -correlation analysis demonstrates that 0, 1, 0 and 0 month time delay separately for the clusters I, II, III and V exists. It illustrated the influences of the water loading and pore pressure diffusion on induced earthquakes. Cointegration tests and impulse response analysis indicates that the 175 m water level only had long term and significant effects just on the seismic events in the intersection region of the Fairy Mount Fault and Nine-brook Fault. One standard deviation shock to 175 m water level increased the seismic activity in cluster Ⅴ for the first-3 months, and then the negative influence was shown. After 7 months, the negative impulse response becomes stable. The long-term effect of the 175 m water impoundment also proved the important role of pore pressure diffusion in RIS with time.