The meat colour is the most important quality factor at the moment of the purchase since the consumer uses it as indicator of freshness. Numerous studies on the shelf life of the meat use the triestimulus coordinates of the CIELab space of colour. Although these variables are very useful to typify the meat colour and are objective; the use of these ones raise concern about how the human beings really see the meat. In addition, it is complicated to define the time in which it will be rejected by the consumers. There were used 24 calves of the commercial category yearling calf. There was extracted the longissimus thoracis of every animal, steaks were placed on modified atmosphere packages. For 8 days the meat was evaluated subjectively (1-dark red or brown to 5- extremely brilliant red) and the instrumental colour of the same one measured up. Several technologies (skills) uni - and multi- variate, linear and not linear and an algorithm of artificial intelligence were used to study the relationships among the visual subjective evaluation and the variables of instrumental colour. The use of machine learning and non linear rules give us a more accurate estimation to the note of the assessors than other methods and demonstrates that it is not advisable to interpret the colour display of the meat of a linear form or using one variable. Lightness and chroma are more precise than redness and yellowness indexes.