Even though FOPDT(First-Order Plus Dead-Time) process is most widely applied in the industrial control field, it is difficult to figure out a in precise process model because of the long dead-time problem. Also, control performance may be deteriorated due to the mismatch problem of plant and model. Thus, the accuracy of process identification is the most important problem in FOPDT process control. In this paper, the proposed method using real-coded genetic algorithm outperforms the existing estimation methods that use step-test responses. The proposed strategy obtained useful result through a number of simulation examples.