Abstract: In order to improve the accurate measurement and control of grain situation in the process of grain storage, this paper carried out storage test on corn based on self-built test bin and established a temperature prediction model applied to corn storage process by using random forest method. The results showed that the accuracy of the first random forest training set was about 0.99, the test set was about 0.66, the accuracy of the second random forest training set was about 0.99, the test set was about 0.58. The training set of the third layer random forest model is 0.99, and the test set is about 0.38. The accuracy of the fourth layer random forest training set is about 0.99, and the test set is about 0.95. According to the preliminary analysis of the random forest model, over-fitting, high generalization error and high variance exist in the first, second, third and fourth layers. By optimizing the model parameters of each layer, the prediction model of corn stack temperature based on random forest was finally determined. The results showed that the random forest method performed well in both the training set and the test set. In the process of large sample data processing, the model is more accurate, and provides an important basis for accurate measurement and control of grain storage situation.