The inter-symbol interference in the nonlinear time-varying channel is a serious problem in the wireless communication. In order to overcome it, wavelet neural network equalizer using error feedback is employed to cut the auto-correlation of the error signal. Exploiting the decent time-frequency localization of the wavelet analysis, as well as the self-training feature of the neural network, a quicker convergent nalysis and computer simulation confirm the effectiveness of the equalizer and a lower BER are attained. Theoretical aalgorithm. It at The wavelet neural network equalizer based on error feedback advances the communication in the nonlinear time-varying channel.