This paper presents a multiobjective approach to thermal units generation dispatch. The competing objectives are minimization of the cost of delivered power and minimization of environmental and ecological impacts. A mathematical model of 24-hour load allocation problem is developed. The model is dynamic in nature due to the ramp-rate constraints imposed on the thermal units generation. A new multicriteria optimization algorithm, which is based on compromise programming, is proposed. The algorithm is applied to a real dispatching problem with 26 thermal generating units. Sensitivity analysis treating the variations of criteria weighting coefficients is given.