Coffee is basically consumed for the pleasure given by its taste and aroma, that is, the quality and acceptance of the coffee beverage are directly related to its sensorial characteristics. Thus, nowadays coffee quality is basically evaluated by sensorial analysis. However, together with this kind of analysis, it should be important to have available more objective chemical methods to assess coffee quality. One possible approach could be based on the analysis of chlorogenic acids (CGA), since they are considered precursors of coffee flavour and pigments during roasting. In the present work, high performance liquid chromatography (HPLC) analysis of chlorogenic acids was applied to six different Brazilian arabica green coffee samples which were previously characterised by sensorial analysis. The results showed the potential to correlate the chemical data, evaluated by the Principal Components Analysis (PCA) statistical method, with sensorial analysis in order to discriminate the quality of the samples. It was observed that the 3,4-dicaffeoylquinic (3,4-diCQA) and 3,5-dicaffeoylquinic (3,5-diCQA) isomers are very important for grouping the coffees into good and bad samples.