Three models of different level of complexity are detailed in order to predict the cold production of a solar cooling device. The transient model consists on a heat and mass transfer model of the absorber, coupled with a solar collector model which conceives the ambient temperature as variable along the day and comprehends the evolution of the solar radiation throughout the day according the locations of the prototype to be analyzed in addition to the ideal models of the condenser and evaporator. The Equilibrium and Universal Model are both lumped models in which the spacial distributions of the variables and the heat and mass transfer resistance in the system are disregarded. The advantages and disadvantages of each model are discussed.The results of the models are compared with experimental data obtained from four methanol/activated-carbon adsorption refrigerator prototypes. The cold generated is underestimated 12%–20% by the transient model, while it is overestimated 22%–46% by the equilibrium model and 53%–66% by the universal model. Even when the transient model allows some improvements, is a useful tool to analyze the impact that different variables of a solar adsorption refrigeration system have in the generation of cold. The simpler models can be used for preliminary design and feasibility studies of solar adsorption refrigeration devices (SARS) whose results can be extrapolated on the premise of its overestimation of cold generation approximately given by this work.