Modeling of rotary drying remains a challenging research topic due to the highly nonlinear and strongly interactive multivariable process. The first-principles model (white box) consisting of partial differential equations with several experimental parameters is very complex and the data-driven model (black box) is lack of transparency and the required depth and quality of industrial experimental data for model training is sometimes difficult to obtain. Therefore, a novel SVR-based hybrid modeling method applied to rotary drying is presented in this paper. First, a first-principles model is set and the unknown parameters of the model are estimated by using a SVR model, then an equivalent fuzzy rule is obtained from the SVR model to improve the transparency of model. The proposed SVR-based hybrid modeling method has been applied to a rotary dryer. The results indicate that the SVR-based hybrid model has better adaptation and prediction capabilities than its black box model.