Objective:To quantify the severity of unconsciousness with (electroencephalography,EEG) nonlinear dynamics analysis of approximate entropy (ApEn) and cross-approximate entropy (cross-ApEn) and to investigate the regularity of changes of EEG nonlinear properties. Method:The patient group included thirty unconscious patients with severe brain trauma or stroke. All of them were diagnosed as unconsciousness with clinical and electrophysiological assessment. Thirty brain injuries patients with normal conscious state were acted as control group. EEG were recorded in all subjects under three states: eyes closed, nociceptive stimulus and auditory stimulus. EEG nonlinear indexes of ApEn and cross-ApEn were calculated for all subjects. Result:EEG nonlinear indexes in patient group were significantly lower than that in control group under all states. In control group, there were significant differences for different states with ApEn and cross-ApEn indexes. Compared with eyes closed state, ApEn and cross-ApEn indexes were higher under nociceptive and auditory stimulus state. In patient group, there was no significant difference for different states with both nonlinear indexes. Conclusion:With EEG nonlinear dynamics analysis of ApEn and cross-ApEn, the degree of suppression of brain cortex in unconscious patients could be directly measured. EEG nonlinear dynamics analysis of cross -ApEn is a more promising method for investigation of consciousness (especially for the level of awareness) because it can reflect both temporal and spatial changes of brain function. EEG nonlinear dynamics analysis could provide information about awake and awareness level of consciousness.