Our research is based on brain-computer interface controller for the realization of the brain communicate with the device. In this paper, we study how to obtain the characteristics of the eeg signals, during the experiment, to acquire EEG signal through the E-Prime psychology software and Neuroscan company production of 64 channel EEG acquisition equipment. Analysis of the original EEG signals using wavelet algorithm, frequency wavelet coefficients as the feature extraction of interest. Analysis of EEG power spectrum, design a classifier based on Fisher criterion. Based on the study of brain computer interface (Brine-Computer Interface, BCI) basis, the movement control was achieved through the imagination. In wheelchair for the actual control object, only through the two electrode EEG signal, the controller design, the theory is validated by experimental success through the imagination of four direction control the wheelchair movement. Finally this paper also discusses the application prospect of brain-computer interface.