Abstarct: One of the important parameters for choosing a good gas sensor, is its selectivity over a particular gas; This means that it can detect a specific gas in an environment filled with several gases. Due to the overlapping of sensor responses to different gases, it is almost impossible to know exactly what gas is in the environment. Several ways have been proposed to solve this problem, one of them is using of a sensor array in which all the components used are not completely selective so it responds to several types of gas. Finally, different methods of data classification can be used to identify the type of gas available, whose response data is collected by a sensor or an array of sensors. In this research, one of these methods has been used as a support vector machine for this purpose. The sensor array was trained for ethanol and carbon dioxide and then tested for the data from the mixture of these gases. This has been repeated in different condition untill the best possible accuracy was reached.