This work focuses on modelling chilling machines based on the absorption-desorption of gas on a solid matrix. In such systems, the cold production changes cyclically with time due to the batchwise operation of the gas/solid reactors. The accurate simulation of the dynamic performance of the chilling machine has proven to be difficult for standard computers when using deterministic models. Additionally, some model parameters dynamically change with the reaction advancement. A new modelling approach is presented here to simulate the performance of such systems by using neural networks. Feedforward full connected neural networks with a single hidden layer were found to be a good and fast tool in order to predict the mean cooling power given by a chilling machine under different operating conditions. The backpropagation learning rule and the sigmoid transfer function have been applied in these neural networks.