We analyzed mercury forms in desulfurization gypsum generated at a coal-fired power plant via temperature-programmed desorption (TPD) and cold vapor atomic absorption. The gypsum used as samples was obtained from double contact flow and jet bubbling desulfurization equipment. There were three types of gypsum samples for each type of desulfurization equipment, and six types of samples in all. The total mercury in these samples was 0.36-1.85 ppm. For model samples, we prepared and analyzed carbon with mercury that was adsorbed onto unburned carbon particles separated from fly ash or onto bituminous coal-based activated carbon. The desorption temperature varied depending on the carbon species; as the amount of adsorption increased, the peak width also increased. The TPD behavior of desulfurization gypsum was simulated with a linear combination of TPD curves from model samples. Mercury was found in carbon contained in all desulfurization gypsum analyzed in the present study. Mercury in carbon in the desulfurization gypsum obtained from the double contact flow desulfurization equipment could be simulated with model samples for unburned carbon with adsorbed mercury alone. Further, desulfurization gypsum obtained from jet bubbling desulfurization equipment could be simulated using unburned carbon with adsorbed mercury and activated carbon with adsorbed mercury. With the jet bubbling type of equipment, the abundance ratios of HgO, HgS, and HgSO4.2HgO were higher than those in double contact flow equipment, showing that depending on the type of desulfurization equipment, not only the type of mercury in carbon but also mercury forms in gypsum vary notably.