According to meteorological element data of test reference year (TRY), a dynamic simulation program calculates the hourly cooling loads of an office building from April to September. Then, a general Visual Basic program is developed based on the error back-propagation (BP) algorithm of artificial neural network (ANN). The network is trained and tested by the obtained data. The results are presented and discussed. The results show that the predicted data is in good harmony with the calculated data, which indicates artificial neural network is a novel and reliable method to predict cooling load.