Brain Computer Interfaces (BCIs) are the systems that enable users who lost their motor capabilities due to neuromuscular diseases to communicate with their environment through the analysis of brain activity. P300 event related potential is one of the widely used signals in BCI applications. In this study, it is aimed classification of P300 potentials by using Discrete Wavelet Transform (DWT) and Linear Discriminant Analysis (LDA) techniques. The proposed method is validated on BCI Competition III P300 dataset provided by the Wadsworth Center. The features that are extracted by wavelet transform showed significant differences between target and non-target stimuli. According to the classification results, 58% and 93% character prediction accuracy is achieved for 5 and 15 intensifications, respectively.