The study proposed an automated sleep-wake scoring system by using the features of electroencephalogram (EEG) and electromyogram (EMG). The method of normalization has been used to reduce the difference of features between the subjects. The proposed system would automatically discriminate the sleep-wake states into both three-state (waking, NREM and REM) and five-state (waking, NREM stage 1, NREM stage 2, transition sleep, and REM). The automated scoring results were compared with the manual scoring results that scored by the experts. The results of global agreement between the automated scoring and experts consensus in three-state scoring are 92.5% (κ=0.88) and 95.3% (κ=0.91) in five-state scoring. The results indicated that the performance of the automated sleep-wake staging system has highly reliability and accuracy.