Abstract: The multi-time effect dynamic monitoring of the total grain yield and its geographical distribution in China was made by use of the sea temper-ature data in the Pacific (286 grid points) in this paper.In order to pred-ict once the whole yield field, the expansion of empirical orthogonal funct-ion was made for initial temperature field，the first several principle compon-ents were drawn to make multiple correlation calculation of sea temper-ature grid point，in order to select predicting factors and construct predicting equation, The prediction model was adopted as dynamical chain in which the time effect could be appointed randomly.It can be used to monitor con-tinuously the total grain yield in China with the sea temperature data.

Key words: Yield prediction, Sea temperature, Dynamic monitor