Abstract: A class of single-input-single-output(SISO) uncertain nonlinear system is concerned. An on-line state estimate algorithm is presented based on a wavelet multiscale analysis. The basic idea is, for the nonlinear system, to perform linear deployment first, then unknown parameters are identified with the application of suitable wavelet functional tests, and finally the best optimizing state estimate of a Kalman is adequately compensated by results obtained. Simulation results are given to demonstrate the effectiveness of the newly proposed method.

Key words: nonlinear systems, multiscale analysis, the best optimizing state estimate