Water loss is a critical problem affecting the quality of table grapes. Temperature and relative humidity (RH) are essential in this process. Although mathematical modelling can be applied to measure constant temperature and RH impacts, it is proved that variations in storage conditions are normally encountered in the cold chain. This study proposed a methodology to develop a weight loss model for table grapes and validate its predictions in non-constant conditions of a domestic refrigerator. Grapes were maintained under controlled conditions and the weight loss was measured to calibrate the model. The model described the water loss process adequately and the validation tests confirmed its predictive ability. Delayed cooling tests showed that estimated transpiration rates in subsequent continuous temperature treatment was not significantly influenced by prior exposure conditions, suggesting that this model may be useful to estimate the weight loss consequences of interruptions in the cold chain.