To meet the government’s demand of developing characteristic agritourism, the agrometeorological observation data of oilseed rape in Gaochun, Nanjing. It is found there is significant correlation between oilseed rape florescence and effective accumulated temperature of t equal to or higher than 3 ℃. By using an optimal correlation analysis method, the optimal factors that affect the model performance are then extracted to establish a mid-term meteorological predicting model for florescence and two long-term predicting models for florescence based on atmospheric circulation and SST, the confidence level α of which both reaches 0.001, and the oilseed rape florescence can be predicted 10 to 30 d in advance. The inspection shows the three models perform well in fitting results and experimental prediction in the past three years. The multiple correlation coefficients of the three models are used as weighting coefficient to calculate weighted mean and the ensemble prediction performs better. The prediction of oilseed rape florescence has practical meanings for people's spring outing and government's investment attraction, and can provide reference for predicting the florescenice of other crops.