In this study, we proposed a novel signal-to-signal translation method based on StarGAN, which generates artificial EEG for steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs). The proposed model was trained using three subjects’ EEG data. The trained model generated artificial SSVEP signals using 15 subjects’ resting EEG data. The probability of improving SSVEP classification accuracy using the generated artificial signals was investigated. We used various SSVEP classification algorithms for the verification like filter bank canonical correlation analysis (FBCCA), combinedCCA, and extension of combined-CCA (combined-ECCA) that we proposed in this study. Using combined-ECCA and our proposed signal-to-signal translation method had the highest performance in terms of classification accuracy and information transfer rate (ITR).