We examined the efficacy of a new method to reduce cybersickness. A real-time cybersickness detection system was constructed with an artificial neural network whose inputs were the electrophysiological signals of subjects in a virtual environment. The system was equipped with a means of feedback; it temporarily provided a narrow field of view and a message about navigation speed deceleration, both of which acted as feedback outputs whenever electrophysiological inputs signaled the occurrence of cybersickness. This system is named cybersickness relief virtual environment (CRVE). Forty-seven subjects experienced the VR for 9.5 min twice in CRVE and non-CRVE conditions. The results indicated that the frequency of cybersickness and simulator sickness questionnaire scores were lower in the CRVE condition than in the non-CRVE condition. Subjects also showed a higher net increase in tachyarrhythmia from the baseline period to the virtual navigation period in the CRVE condition compared to the non-CRVE condition. These results suggest that a CRVE condition may be a countermeasure against cybersickness.