Municipal solid waste incineration (MSWI) has gradually become the main technology of waste treatment because of its efficient capacity reduction. However, the nitrogen oxides (NOx) produced in the MSWI process are one of the main pollutants. In order to control NOx emissions while ensuring combustion efficiency, an intelligent optimization setting method of air flow for MSWI process based on multi-objective particle swarm optimization is proposed. Firstly, by the combined minimal-redundancy maximal-relevance criterion and the feedforward neural network, the prediction models of combustion efficiency and NOx emission are established. Then, an improved staged multi-objective particle swarm optimization algorithm (SMOPSO) is presented to obtain the Pareto optimal solutions of primary air flow and secondary air flow. In addition, the utility function is designed to determine the optimal setting value of the primary air flow and the secondary air flow. Finally, the simulation experiments verify the validity and feasibility of the proposed method based on the practical operation data.