This article presents the implementation of an Electronic Nose (EN) based on the development of a measuring chamber composed of a matrix of 6 metal oxide gas sensors with partially overlapping sensitivities, to identify and classify the volatiles emitted from the exhaled breath of a group of healthy people and another group with samples of patients with Chronic Obstructive Pulmonary Disease (COPD). A set of algorithms for pre­processing and signal processing based on techniques of feature extraction and statistical methods as Principal Component Analysis (PCA), for pattern recognition of data set were implemented. From the exhalation breath samples of these patients, the results were important, due to the classification of the patients with the disease as well as of the healthy voluntary group.