In an electronic nose, the most important component is the sensor array and the classification accuracy of an electronic nose that depends significantly upon the choice of the sensors in the array. While deploying an electronic nose for a specific application, it is observed that some of the sensors in the array may not be required and only a subset of the sensor array contributes to the decision. Thus, the number of sensors used in the electronic nose may be minimized for a particular application without affecting the classification accuracy. In many cases, the sensor array produces an imprecise, incomplete, redundant, and inconsistent dataset and thus the classification accuracy degrades due to these redundant sensors. The rough set theory is a mathematical tool capable of selecting the most relevant and nonredundant feature from such datasets. In this paper, the notion of rough set theory is utilized for pattern classification in an electronic nose with black tea samples and at the same time optimization of the sensor set is carried out.