This paper focuses on improving thermal efficiency and reducing unburned carbon in fly ash by optimizing operating parameters via a novel high-efficient swarm intelligence optimization algorithm (grey wolf optimizer algorithm, GWO) for coal-fired boiler. Mathematical models for thermal efficiency and unburned carbon in fly ash of the discussed boiler are established by artificial neural network (ANN). Based on the ANN models, the grey wolf optimizer algorithm is used to obtain higher thermal efficiency and lower unburned carbon by optimizing the operating parameters. Meanwhile, the comparisons between GWO and particle swarm optimization (PSO) and genetic algorithm (GA) show that GWO has superior performance to GA and PSO regarding the boiler combustion optimization. The proposed method can accurately optimize the boiler combustion performance, and its validity and feasibility have been experimentally validated. Additionally, a run of optimization takes a less time period, which is suitable for the real-time optimization. INDEX TERMS Coal-fired utility boiler, grey wolf optimizer, thermal efficiency, unburned carbon in fly ash.