A new method is presented for the off-line identification of plant nonlinearities from plant response data. No prior knowledge of the structure of the nonlinearities is required, although a knowledge of the dependent variables within the nonlinearities is assumed. The identification strategy involves the use of Walsh functions which are described in the paper. These functions are particularly suitable for high-speed digital implementation, and are very well suited to dealing with functions with discontinuities. The nonlinear functions identified by this method can be used in a nonlinear plant Walsh function control scheme, which is currently the subject of research. A set of typical implementation results of both identification and control is included, which shows the efficacy of the method.