Electroencephalogram (EEG) signal preprocessing is one of the most important techniques in brain computer interface (BCI). The target is to increase signal-to-noise ratio and make it more favorable for feature extraction and pattern recognition. Wavelet transform is a method of multi-resolution time-frequency analysis, it can decompose the mixed signals which consist of different frequencies into different frequency band. EEG signal is analyzed and denoised using wavelet transform. Moreover, wavelet transform can be used for EEG feature extraction. The energies of specific sub-bands and corresponding decomposition coefficients which have maximal separability according to the Fisher distance criterion are selected as features. The eigenvector for classification is obtained by combining the effective features from different channels. The performance is evaluated by separability and pattern recognition accuracy using the data set of BCI 2003 Competition, the final classification results have proved the effectiveness of this technology for EEG denoising and feature extraction.