Coffee is one of the most consumed beverages in the world and its consumption has increased due to improvements in its quality mainly related to better agronomic practices, association with health benefits and availability of new coffee varieties, which significantly influence on sensory attributes. Nonetheless, sensory attributes demand-trained tasters, increasing the costs for quality analysis in industry. In order to verify the correlation between quality properties and sensory attributes, 186 coffee samples from Brazil (Paraná and Minas Gerais, arabica and conillon) were evaluated according to Brazilian recommendations. The company that processes the coffee provided coffee samples and sensory responses. The evaluation was performed by a least squares analysis considering each quality property and sensory attribute, being the Pearson correlation calculated to each one. The results indicate that the quality properties background granulometry and infusion are no correlated with the sensory attributes, since these quality properties presented Pearson correlation values (disposed on the side scale) equal to zero or below 0.5. The color (L) presented Pearson correlation around 0.7 with: powder fragrance, drink aroma, bitterness, bulk, astringency and final grade. Correlation around 0.78 was observed between color (L) and flavor. The residual flavor was not correlated with any quality property. According to the results, it was verified that sensory attributes, with the exception of residual flavor, could be estimated from color (L) quality properties. Through the results observed using an univariate calibration curve, obtained from the color (L) against each specific sensory attribute (obtained from trained tasters, in this step), it was possible to estimate the sensory attributes to new samples, reducing considerably the industry costs.