The invention belongs to the technical field of hot continuous rolling process control, and discloses a deformation resistance phenomenological model calculation method based on data mining to solve the problems of single parameters and incomplete consideration factors of a traditional deformation resistance model. The deformation resistance phenomenological model calculation method specifically comprises the following steps that a deformation resistance sample library meeting the requirement of model calculation is established and regularly maintained, and steel coil data with large deviationis eliminated; model regression calculation and deformation resistance self-learning coefficient calculation are periodically triggered; deformation resistance phenomenological models are established; regression calculation is carried out on the phenomenological models according to steel grades to obtain model parameters of different steel grades; according to the model parameters after regression, the corresponding self-learning deformation resistance coefficients of various frames can be calculated; the model parameters of different steel grades and the corresponding deformation resistanceself-learning coefficients of the various frame are sent to a process control system. The deformation resistance phenomenological model calculation method can obtain effective data and the accuracy ofregression and model calculation are improved, so that the product quality is improved and the yield is increased.