Heated air by solar collector has high variation in temperature because it is depend on time. Therefore, a rock-bed was used to regulate the air temperature and to damp out the large fluctuation of air temperature, although it causes a change in the phase and the amplitude temperature of the inlet air. The rock-bed could be used then to store day-heat for night use. The purpose of this paper was to mathematically model the rock-bed performance using SAS language. This modeling utilizes; (1) rock-bed characteristics, (2) Fourier series, and (3) statistical procedures for predicting the outlet air temperature. A computer program was developed to achieve this purpose. An air flow of 4.66 m 3 /min was used to test the computer model and air temperatures at 0, 0.375, and 0.70 m of the rock bed height were recorded by time for 24 h. It was found that Fourier series was more efficient for predicting the air temperature at any vertical location of the rock-bed. The correlation coefficient using the rock-bed characteristics and Fourier series was 0.905 for predicting the outlet air temperature. The determination coefficients were found to be above 0.98 between observed and predicted air temperatures at any vertical location of the rock-bed.