Classification models developed by LDA and PLS-DA achieved correct classified rate of 100% and 97.4% in the prediction, respectively. The discrimination could be attributed to different raw materials (mainly brewing water) and elaboration processes employed by the wineries.