Animal welfare and zootechnical performance are compromised when animals are housed in environments which place them outside their thermal comfort zone. However, the identification of thermal stress, when based on air properties, suggests the use of outdated and generic indices. The objective of this work was to develop and validate a methodology for classifying and diagnosing heat stress in production animals based on psychrometric air relations. The model was created for broilers, pigs, dairy cattle, and laying birds, categorized into a total of 21 breeding phases. For each phase, a bibliographic search was carried out for the psychrometric parameters of the air - dry bulb temperature (AT) and relative humidity (RH) - that satisfied the animals' critical and ideal thermoneutral zones. Adding the local atmospheric pressure (AP), the parameters were used to calculate the enthalpy (h), resulting in five comfort ranges. Based on this, a decision tree was elaborated, consisting of three attributes (AT, RH, and h) and seven diagnostic classes, based on the psychrometric principles of air. The proposed methodology was used in a case study, with a database extracted from an individual shelter for calves. For the evaluation of the decision tree, two induction algorithms, ID3, and c4.5, were compared, both of which presented high accuracy and proposed simpler tree models than the one theoretically developed for the methodology. In conclusion, the methodology represents a great potential to characterize the thermal comfort of the animals, diagnose the causes of stress and recommend possible corrective actions. The study revealed that decision trees can be adapted and simplified for each creation phase.