In this paper, in order to explore underlying the interaction mechanisms between brain regions, the cross entropy measures: cross sample entropy (C-SampEn) and cross fuzzy entropy (C-FuzzyEn) were used to measure the brain synchrony through analyzing the background electroencephalograph (EEG) signals in Alzheimer's disease (AD) patients. It was demonstrated that the values of both the two features in AD were lower than that of the normal controls, particularly in the alpha band, indicating that the synchrony strength and the information transferring efficiency between the corresponding brain regions were decreased in AD. Moreover, classification results showed that the two cross entropy features could classify AD Group and Control Group correctly. However, the C-FuzzyEn performed better, reflected in the higher classification accuracy. Thus, C-FuzzyEn revealed the abnormalities of brain activity for AD by evaluating the synchrony from EEG series, and it seemed to be an important auxiliary clinical method to diagnose AD early.