Aiming at the severe inter-symbol interference and high bit error rate in short-wave fast time-varying channels,this paper designs a short-wave channel blind equalizer based on Convolution Neural Network(CNN),and analyzes the influence of parameters in CNN structure on channel equalization,such as the number of convolution layers,the depth of convolution layer and the size of the convolution kernel layer.By simulating two typical short-wave time-varying channel,Rayleigh flat fading and frequency selective fading channels,we have the following results:1)Compared with the Recurrent Neural Network(RNN)structure equalizer,the CNN has higher accuracy during the training process,the convergence speed is faster,and the stability after convergence is higher.2)Under the condition of simulation,the CNN-based short-wave channel blind equalizer designed in this paper can effectively extract input signal when using 2×3×3 convolution kernel size and 2-layer convolutional layer.The characteristics of the classification layer improve the equalization performance while reducing the complexity of CNN structure.3)For the short-wave channel,the error rate of Convolution Neural Network Equalizer(CNNE)is lower than that of Recurrent Neural Network Equalizer(RNNE)under the same SNR.