This paper proposes the model identification of a DC generator using the particle swarm optimization (PSO) which is one of the efficient metaheuristic search techniques. A DC generator in the engineering laboratory is tested to collect the useful data for identification process. Such the data are step input voltage and step output responses. For comparison, the genetic algorithm (GA) and the tabu search (TS) are conducted for identification. As results, it was found that the PSO can provide very satisfactory mathematical model of a DC generator with more accurate than GA and TS.