A number of works have been reported on robot control using EMG signals. Control of robots, wheelchairs, and rehabilitation aids using the arms, hands or legs by EMG signals has been quite popular and effective. However, few works have dealt with head-movement control using neck EMG signals. We have built a model that estimates continuous human head movement from neck EMG signals. Our proposed model, which considered not only static but also dynamic effects, effectively suppressed the over/undershoot, and predicted head-rotation movements well. This result indicates that the proposed model has the potential to reconstruct the observed data from neck EMG signals properly.