A simple technique for recursive identification of the Wiener-Hammerstein model with extension to the multi-input single-output (MISO) case is presented. We use a new transformation of the input-output difference equation where parameters to be estimated are those of each subsystem of the initial and unique realization. After that, a weighted extended least squares (WELS) method is employed to estimate recursively and separately parameters of the linear subsystems and the static nonlinear element. Convergence analysis of the proposed procedure is also studied. Finally, a numerical example is provided to show the efficiency of the algorithm;