Abstract: To study the problem of the real-time data processing in communication system,a support vector machine equalizer with convergence speediness capability is presented.Based on the technology of support vector machine,train samples are constructed with decision feedback signals,and a wavelet kernel function is adopted and the flex parameter is adaptively adjusted in this equalizer.Then an adaptive equalizer with tracing abilities is designed.This novel equalizer is simulated in software and is compared with the support vector machine equalizer with Gauss kernel function.The result shows the proposed equalizer has much better performance in convergence speed.