Aiming at the secondary air flow of waste incineration process are usually set according to manual experience, which is subjective and arbitrary, so that the pollutant emission concentration does not meet the standard, an intelligent optimal setting method is proposed. Firstly, a case-based reasoning pre-set model, and an evaluation and learning model of secondary air flow setpoint are constructed. Secondly, a stochastic configuration network process index prediction model is established. Then, an intelligent compensation model based on the RBF neural network self-learning fuzzy inference is constructed. Finally, the pre-set model, process index prediction model, intelligent compensation model and evaluation and learning model of setpoint are organically integrated, the structure and function of the intelligent optimal setting method are designed, and algorithm implementation is given. The experimental results based on historical data of a waste incineration plant show that the fluctuation degree of setpoint obtained by this method is less, and the control system running according to the setpoint can reduce the pollutant emission concentration, which can promote the realization of operation optimal goal in the incineration process.