The PH value of slurry in the absorber is an important parameter affecting the efficiency of wet flue gas desulfurization system in coal-fired power plants. The wet flue gas desulfurization system of coal-fired power plant has the characteristics of large lag, nonlinear and strong coupling. It is difficult to accurately control the PH value of the slurry in the absorption tower. In this paper, the advantages of gated recurrent unit (GRU) neural network in processing time series data are used to predict the PH value of slurry in the absorber. Firstly, the data collected in coal-fired power plants were analyzed to screen out the variables with strong correlation with the PH value of grout. Then, the time series data of these variables are used as the input of the model to train the model, and the prediction model of the slurry PH value in the absorption tower in the wet desulfurization system is obtained. Finally, the real data are collected to test the model. The results show that compared with BP neural network model, radial basis function neural network (RBF), recurrent neural network (RNN) and long short-term memory (LSTM) neural network, the proposed model is more accurate and practical.