Mathematical models describe systems using mathematical concepts and language, which can be used in food engineering to describe degradation processes. The essential purpose of kinetic models is to sufficiently describe a set of experimentally obtained data. These kinetic models can be used for prediction, process control, optimization and simulation of various food processing operations. Kinetic models can be used to describe changes in quality attributes in fruits and vegetables, such as firmness, tonality (Hue angle) and weight loss. This repository contains experimental data obtained using a Central Composite Rotational Design (CCRD) with 2 independent variables (temperature and relative humidity). Data on storage firmness, Hue angle and weight loss of 'Pedro Sato' guavas stored at different temperature and relative humidity conditions were used to verify the prediction capacity of semi-empirical kinetic models available in the literature and new semi-empirical models that were modified in the term referring to the metabolic part, with the aim of improving the performance of the semi-empirical models available in the literature and subsequently, applying the models with the best performance in systems capable of estimating the shelf life of guavas during the chain of cold.