Accurate acquisition of flue gas environmental protection system performance is the premise of its optimal adjustment and technological transformation. Based on the limestone-gypsum wet sulfur removal system of a 600 MW coal-fired power station, the data mining was used to obtain the key factor set of desulfurization system performance. On this basis, a variety of algorithms of regression analysis technology were combined to study and construct the data model of the performance of the flue gas environmental protection system of coal-fired power station. The results show that the data model can reproduce the desulfurization system performance under different load conditions with high precision, and also has good flexibility and expandability, which provides a solid foundation for state monitoring, trend analysis and operation optimization of the environmental protection system.