We propose an EEG-based discrimination method for the right/left hand movement in a single trial. The EEG was recorded during the voluntary movement and imagination of the hand movement. We made a feature vector for every second that represents the characteristics to reflect the process of the right/left movement. It was composed of the ERD, ERS patterns of the mu and beta rhythm and the coefficients of the autoregressive model best fitting for the data of the given period. Linear discrimination of their distributions in the vector space classified the right/left hand movement-related EEG activity efficiently.