Abstract: Parkinson's is a neurological disorder that affects the speed of the human motor organs. Early diagnosis of this disease is important to prevent its progression. The efficiency of SVM classification is directly dependent on the selection parameters and the training data. The increase of training data has nothing to do with the accuracy of classification. In this paper, a new approach using the integration of genetic algorithm and SVM classification, is used to diagnose Parkinson's disease. Having used the genetic algorithm, the best training and possible values are selected for the SVM parameters. Eventually, it has been shown that the accuracy of diagnosis of Parkinson's disease using new algorithms has reached 100%.