A new method of reaction rate modeling of volatile kiln was proposed. First, use the collected material mass and kiln temperature data to get the corresponding reaction rate data by experiments, then proceed the data normalization processing and grouped into training samples and test samples, after that, adopting cross validation method to obtained the parameters of support vector machine regression (SVR) model by use of the training sample. volatile kiln reaction rate model was built based on these parameters, and finally use the test sample to verify the accuracy of the model, and the results of the correlation coefficient reached more than 0.9, which shows that the method of SVR's generalization ability is stronger and this reaction rate model can be used as the volatilization kiln reaction rate model.