Hydrogen sulfide (H2S) presents many hazardous traits such as corrosive, explosive, toxic, and flammable. It is slightly denser than air, and a mixture of H2S and air can be volatile. Therefore, a reliable and robust measurement system is required to effectively detect and quantify H2S in many applications, such as oil and gas industries. There are several methods available in the literature to quantify H2S in fuel gases; however, only a few are available in case of air samples. Furthermore, array-based sensors are more reliable in the detection of volatile organic compounds (VOCs); however, sensor arrays are more expensive and challenging to produce. To overcome the limitations of producing physical sensor arrays, this article proposes a concept of virtual sensing that enables to augment a single sensing platform into a virtual array, thus, increasing the detection accuracy at no extra cost of producing a large physical sensors array. The merits of the proposed system are as follows: 1) a virtual sensing concept is combined with a physical sensing platform to enhance the proposed model’s estimation power in quantifying H2S in air samples; 2) a new feature extraction method based on fractional derivatives is proposed to further enhance the model’s learning capabilities; 3) an array of four gas sensors is fabricated in the in-house foundry to record and analyze the signature of H2S at various concentration levels; 4) a shallow neural network (NN) model is trained and tested on the recorded data, and based on the NN’s input–output relation, a mathematical model is presented for the quantification of H2S; and 5) the proposed model is a highly sensitive and reliable H2S gas sensing scheme with the ability to detect the gas instantaneously. The proposed gas quantification model has the advantages of being low cost, easy to implement, and fast operation compared with the analytical solutions. Furthermore, it is extensively tested and validated using real gas data. [ABSTRACT FROM AUTHOR]