Multisensor arrays constituted by non-specific sensors, like in the case of electronic noses, is an appropriated approach to identify the presence of a particular gas in a mixture. While the dependence of the sensors signal output by the concentration of each single gas can be determined through a sensor calibration procedures., it is much more difficult (and of non- practical use in most of the cases) to know the response of the individual sensors when exposed to mixtures containing several components. In a previous work, Bartlett and Gardner have used a phenomenological linear model to simulate the response of semiconductor sensors to such a mixture. The concentration dependence of a sensor signal in a gas mixture of N components of concentrations {C1,…,CN} was calculated as a linear combination of the sensors response to every individual gas in an artificial and well controlled atmosphere. The aim of this work is to analyze the capability of an electronic nose model to identify gases, mainly CO and NO₂, normally present in polluted atmospheres. It is relevant also to identify organic reducing gases like isobutane (ISBU) which is emitted in a variety of different industrial processes. To reach the objective it is not a simple task because the output of each individual sensor is non-specific, but conditioned by the presence of the different gases. For example it is well known that tin dioxide (SnO₂) sensors are particularly sensitive to alcohol's, mainly ethanol (EtOH). Other gases, although not pollutants, such as methane (CH₄), are present in the atmosphere and its concentration can drastically increase due to emissions from gas lines or natural fermentation processes. For these reasons we have implemented a theoretical electronic nose model, composed by a few number of commercial tin dioxide sensors (Taguchi), addressed to identify the presence of CO, ISBU, CH₄ and EtOH and to determine their concentration. In this case the gas concentration ranges between few hundred to thousand parts per millions, which is the characteristic for Taguchi sensors. The concentration dependence of the sensors signal output to the single gas was taken from available Figaros Inc. sensor data sheets. Efforts have been made not only to identify the individual gas in humidified synthetic air, but to obtain the gas concentration of each individual compound, Ci, when the gases are contemporary present in a mixture of a given total analytical concentration. The linear response dependence assumed by Bartett and Gardner was generalized using non-lineal interpolation functions. In a second part of the work, the electronic nose model was simulated using the experimental output of 100 nm thick SnO₂ sensors deposited upon a substrate heater element realized upon micromachined silicon substrate having 200 nm Si₃N₄ membrane as physical support of the entire sensor stack and developed by the LAMEL Institute, Bologna. The sensitivity towards several reducing and oxidizing gases is in the range of small fraction of ppm. This sensitivity level makes them suitable for environmental monitoring of pollutant gases such as CO, NO₂, benzene, and a mixture of toluene and xylene (TX).