Due to the increased use of medicinal plants, the qualitative classification is inevitable. Rosa damascena Mill. with a high value of essential oil and its unique properties in the health, food and pharmaceutical industries is of one of these plants. In this study, after essential oil extraction from nine genotypes of Rosa, the essential oil components were identified by GC and GC-MS analysis. The genotypes were divided in three classes (C1, C2, C3) based on total percentage of six most important compounds, having major role in essential oil quality (phenyl ethyl alcohol, trans rose oxide, citronellol, nerol, geraniol, geranial).Then, the classes were tested by an electronic nose (EN) system designed based on metal oxide semiconductor (MOS) sensors. Sensors response pattern was recorded and analyzed by chemometrics methods in next step. Results of principal components analysis (PCA) showed that 85% of data variance was explained by two first principal components (PC1, PC2). Artificial neural network based on back propagation multilayer perceptron (Bp-MLP) was performed and classification accuracy achieved 100% and 96% for training and test sets, respectively. These results showed that EN could be used as a quick, easy, accurate and inexpensive system to classify Rosa damascene Mill essential oil.