Abstract: Spatial model from perspective，the study performed the Moran′I computerization to analyze the spatial autocorrelation characteristics of the oilseed yield on country level，based on the data about oilseed yield of 87 country units in Shandong province.Results showed that the Moran′ I coefficient 40 country units with oilseed yield has obvious clustering characteristics，the autocorrelation of the other country units was not evident.On the other hand，the study also carried out to predict the total oilseed yield with ARMA（2，1，2）model on basis of time series data，in order to explore the trend of the total oilseed yield in Shandong province，the average relative error between observation value and prediction value is 2.12% only using statistical oilseed yield data，which were gathered annually over the course of 31 years between 1978 and 2008.In a word，Moran′ I coefficient and ARMA（2，1，2）model can fairly clarify the spatial and temporal status of oilseed yield.What′s more，the study is to provide a better understanding of spatial and temporal patterns of oilseed yield in Shandong province.