Abstract This paper deals with the discrimination of gases using the responses of an integrated gas sensor. The multiple regression method is used to find the sensor parameters for different gases in training mode. In sniffing mode, these parameters are used to discriminate the gases present in the ambient using the iteration technique. These methods, i.e., the multiple regression method and iteration technique, have been successfully applied to the data set obtained in our laboratory for a mixture of carbon tetrachloride and ethyl methyl ketone exposed to a fabricated integrated gas-sensor array. The methods can effectively be used for the realization of an 'intelligent gas sensor'.