Abstract: In view of the significant impacts that polar ice has on the climate of the mid-low latitudes,this paper mainly analyses the effects and possible mechanism of Arctic sea ice on the precipitation in the main flood season of the Yangtze River basin by statistical methods such as Singular Value Decomposition. The precipitation data is collected from 116 rainfall stations in the Yangtze River basin and the sea ice data is provided by the United States National Snow and Ice Data Center (NSIDC),both periods are 1979-2017. The results show as follows:In winter and spring,when ice area in the Barents Sea and Okhotsk Sea was more than the average and that in the Beaufort Sea was less than the average,then summer precipitation in upper and middle reaches of the Yangtze River,the upper reaches of the Hanjiang River and the Yalong River increased;When winter-spring ice area in the northern part of the Arctic Archipelago,Chukchi Sea and Laptev Sea was much,summer precipitation in the Dongting Lake and Poyang Lake was much too,but it was less in the upper reaches of the Jialing River and Hanjiang River; and vice versa. The possible mechanism is that the sea ice variation in key regions in winter and spring stimulates atmospheric energy fluctuation by affecting turbulent heat flux. This fluctuation propagates to East Asia in the form of atmospheric wave trains and affects the summer atmospheric circulation and water vapor transport in East Asia,thus indirectly affecting precipitation of the Yangtze River basin in the main flood season. Based on the multivariate regression method,four prediction models of precipitation in the main flood season are built by taking the sea ice area in key region as predictors. The models have obvious fluctuation on quantitative precipitation prediction,but it has a relatively good effect on precipitation trend prediction.