This paper puts forward a new viewpoint on optimization of boiler combustion, namely, reducing NOx emission while maintaining higher reheat steam temperature rather than reducing NOx emission while improving boiler efficiency like traditional practices. Firstly, a set of multioutputs nonlinear partial least squares (MO-NPLS) models are established as predictors to predict these two indicators. To guarantee better predictive performance, repeated double cross-validation (rdCV) strategy is proposed to identify the structure as well as parameters of the predictors. Afterward, some controllable process variables, taken as inputs of the predictors, are then optimized by minimizing NOx emission and maximizing reheat steam temperature via multiobjective artificial bee colony (MO-ABC). Results show that our rdCV-MO-NPLS model with MO-ABC optimization methods can reduce NOx emission synchronously and improve reheat steam temperature effectively compared with nondominated sorting genetic algorithm II (NSGA-II) and combustion adjustment experimental data on a real 1000 MW boiler.