In the present work an artificial neural network (ANN) model is introduced which was elaborated for modelling the layer temperatures in a storage tank of a solar thermal system. The model calculates the temperatures of 8 equal layers of the storage tank in several time interval from the average time interval data of the solar radiation, the water consumption, the ambient temperature, the mass flow rate of collector loop and the temperature of the layers in the previous time-step. The used time intervals are one hour, 30, 10, 5, 2 and 1 minutes. The introduced ANN model is convenient for describing the system in every case, and the identified models give acceptable results inside the training interval. The average deviation was 0.53°C during the training and 0.76°C during the validation in case of hourly data and these data were 0.07°C and 0.08 in case of 1 minute time interval. The optimal time interval was found at 5 minutes.