Sulfide precipitation of nickel and cobalt was investigated in three reactors: batch, seeded fluidized bed with an aqueous sulfide source, and bubble column with a gaseous sulfide source. In terms of metal removal and settling rate, the sulfide system is more efficient than either hydroxide or carbonate. In the batch system, the kinetics of the nickel and cobalt sulfide precipitation reactions are first order in metal and sulfide concentrations. Metal removals of approximately 90 % were achieved, but when excess sulfide was used the removals decreased, probably due to the formation of aqueous polysulfide species. In the fluidized bed, significant fines were measured, due to homogeneous nucleation and attrition. It was concluded that this reactor configuration is not suitable for these very high supersaturation levels. Using a gaseous sulfide source reduced the supersaturation and the measured metal removal improved.