The rape is one of the most major oil plants. It is important to forecas t the yield of rape dynamically for agricultural production and national foodstuff secur ity. In this study, a dynamical prediction method has been established both for regionspecific rape yield and nationwide yield. The method utilizes the data o f rape yield, development stage, physiological meteorological index, daily maximu m and minimum temperature, daily precipitation, and daily sunshine duration from individual major producing region and the whole china. The method involves with the calculation of composite diagnostic index from the correlation coefficient and Euclidean distance and the derivation of meteorological influence index for bumper or poor harvest in the prediction year from the historic meteorological i nfluence index. The results indicate that the method has high prediction accurac y of the bumper or poor harvest trend and the actual yield of rape on both trial and test of forecasting and can satisfy the needs of operational application. T he method not only can predict efficiency in advance, but also realize the capab ilities for dynamical tracking and prediction and has great potential in operati onal application. 