A sEMG signal is useful for the control of the wearable robots because its measurement is relative-easy and the joint torque can be estimated from sEMG signals of related multiple muscles. However, estimation methods of sEMG signals are not always available for all users. On the other hand, an EEG signal is one of the biological signals which might be the substitute for a sEMG signal because an EEG signal can be measured even with persons who cannot provide some sEMG signals. In general the estimation of the user's motion-intention from EEG signals is more difficult than the estimation from sEMG signals. The robot might work against the user's motion-intention when the user does not concentrate on the control of the robot and is distracted by other things. In this paper, to estimate user's motion-intention from EEG signals, the feature of EEG signals during upper-limb motion is investigated and we estimate whether or not the user moves the upper-limb.