Abstract: [Objective] In order to reduce the loss of agricultural production brought from extreme precipitation event. [Method] Using the data of daily precipitation data of 17 cities in Anhui, the number of annual extreme precipitation event on 1965 ~ 2014 was counted by the centesimal value method. On this basis, the temporal-spatial distribution characteristics of extreme precipitation event (1988~ 2013) were analyzed by using the method of linear trend analysis and EOF analysis, and Based on the structural vector auto regression model in econometrics, take extreme precipitation event as a factor and build a new climate-economy model with output value of agriculture, gross domestic product and agricultural machinery gross power together, and then, using the means of impulse response functions and variance decomposition, the dynamic effect of extreme precipitation events on Agricultural production can be analyzed. [Result][Conclusion] The results show that: in nearly 50 years, 11.59 extreme precipitation event occurred each year on average in Anhui, no interannual variation trend, it showed d a diminishing scale from south to north. There were no interannual variation trends in all regions too. EOF analysis showed extreme precipitation event had two typical distributive characteristics in Anhui, first was coherently change in all regions and mid Anhui changed more than south and north, the center of change was the eastern part of Yangzi-Huai valley. Second was changed opposite, the center of change was southern mountains and center of along and north of Huaihe River. Both of two distributive characteristics appeared oscillatory variation in time, and in nearly 5 years, these two typical distributive characteristics manifest as consistent more occurred and more occurred in south, less occurred in north. The overall effect of extreme precipitation event on agricultural production was positive , the extent of effect is about 0.01 %, but the contributions ratio was quite low.