Abstract: Agriculture is foundation of national economy，and grain is the priority among priorities.Accurate prediction of grain yield is of great significance in national food security and making grain production policy.Using grey correlation analysis method，specific factor in Yunnan grain output was analyzed from numerous fuzzy factors.The factors followed in an order as following：total power of agricultural machinery，fertilizer rate，available irrigation land area，rural electricity consumption prediction.These factors were corrected by GM（1，1）residual correction.Predicted data was used as a sequence of related factors，and grain yield was regarded as characteristic sequence of system.A grey GM（1，N）prediction model for grain yield was constructed.According to grain production related data in Yunnan province from 1999 to 2015，grain yield data of Yunnan province in 2020 was predicted.Fitting data until 2015，average relative error between actual output and simulation value was 1.92%，which has high accuracy.Through analysis of prediction results，suggestions were put forward about stabilizing and improving food cereal production in Yunnan province.