Coal characterization and prediction of slagging behavior has reached great relevance for coal-fired power plant operators. Especially, for units subject to large fuel quality variability, fuel switching, and coal blending. Laser-induced breakdown spectroscopy (LIBS) and neural networks were used to characterize elemental composition of coal samples and estimate the initial deformation ash fusion temperature. A coal inventory was assembled with a range of slagging characteristics and fusion temperatures from 955 to 1480 °C. The following inorganic elements were measured in the laboratory: Al, Ca, K, Mg, Na, Fe, Si, and Ti, and correlated to coal ash initial deformation fusion temperature under reducing conditions. The LIBS system achieved elemental composition measurement accuracy within ±15% (absolute). The LIBS system was tested off-line at a power plant on three different coals. The field results indicate an average relative fusion temperature prediction error when compared to American Society for Test...