In this paper, we consider identification of a nonlinear additive system. An input signal is designed in such a way that the problem of identification of nonlinear additive systems is reduced to a problem of identification of static nonlinear functions. Then, three approaches are established to estimate the order of the system. The methods exploit the structure of the nonlinear additive model so that their implementations are easy. (c) 2005 Elsevier Ltd. All rights reserved.