Functions of human brain can be well analyzed by recording and analyzing the electrical activities of the brain. These analysis are very useful in providing accurate diagnosis and treatment of any disorders of this type. In the diagnosis of any neurological disorders, it is always necessary to give utmost importance to its methods of analysis. Despite poor spatial sensitivity compared to other methods of analysis of brain activities, EEG is used for this purpose because of its low cost recording machines, high temporal resolution etc.Although different works have been reported related to the diagnosis of epilepsy / Parkinson’s disease (PD), we have chosen the application of simple EEG comparison techniques for the purpose. We have taken samples of 10 numbers of subjects suffering from PD and 10 numbers of non PD persons. Prior to the extraction of any features for classification of these EEG signals, it is utmost important to preprocess them. This is necessary in order to remove the artifacts present in the signal considered for processing. In this paper, we are presenting an application of wavelet transforms (WT) by which the noise contained in the raw EEG of persons suffering from PD and of normal conditions can be reduced or eliminated with high efficiency.