Long-term alcohol use can affect brain function. This is what prompted researchers to analyze the electroencephalogram (EEG) signal to recognize EEG signals in alcoholic or non-alcoholic persons. In this study, a method of classifying Alcoholic and non-alcoholic EEG signals using an image recognition approach is proposed. The EEG signal consisting of 64 channels is processed through a Fourier transformation process and arranged into an image-like plot with a 64 × N/2 size from the N-point FFT used. Furthermore, this plot is classified using a Convolutional Neural Network (CNN) to determine whether the EEG signal belongs to an alcoholic or non-alcoholic person. The test results showed the highest accuracy of 98.7% at 512-point FFT. This result is higher than the classification process using EEG signals in the time domain. The proposed method has great potential for EEG signal processing because of its simple computation.