Abstract
This paper considers a procedure of identifying Hammerstein models with white noise input, assuming that only a finite string of output data is available.A linear dynamical system is first identified by using the output data and then a static nonlinear function is estimated from the probability density function .A numerical simulation is also included.