The authors investigate feasibility of application of the charging/discharging effects observed in the gas-sensitive metal oxides under the effect of a longitudinal electric field, and their variations at different atmospheres for a selective analyze of the gas of interest. The authors employ a concept of the gas-analytical devices, which operate on the principles of the biological olfactory system. Such devices, often referred to in literature as "electronic noses", are based on the multi-sensor array, which generates a primary vector signal under the influence of the gas mixture, and a pattern recognition technique based on the algorithms, which process the cumulative multi-sensor array signal. Like in case with a biological analogy, the analysis of the gas mixture is carried out not due toits decomposition into components, as it is done in the traditional spectrometers, but due to an odor "image" recognition complex. For measuring of the resistance of each segment of the multi-sensor array chip, an electric field (potential difference) is applied. The transient response of the measuring current, as it was shown in our previous studies, depends on the kind of a gas and the concentration of its constituent components. Here we employ a chemiresistor array based on the SnO2:Pt thin film segmented by multiple electrodes into a chip. The authors describe the charging processes in such structures experimentally observed at the operating temperatures of about 300 °C. The longitudinal electrical field applied along the metal oxide film is distorted. This electric potential distortion depends on the gas and may serve as an "image" (or a vector signal) of the gas. We process this "image" with the use of a linear discriminant analysis. The results obtained demonstrate feasibility of this method for a selective identification of the gas of interest and development of new gas-detection instruments based on this principle.