The relationship of the rainfall rate averaged over a sampling area, which is sufficiently large and conditionally homogeneous of climate, with the cloud coverage, standard deviation of the brightness temperatures at cloud top,and temporal variability of the cloud coverage, has been studied, and three statistical models are obtained for estimating rain amount with digitized, 1 h interval infrared images from geostationary satellite. The results from the precipitating weather on the 5th, 6th and 10th of July in 1991 show that the models are quite good at the estimation of daily precipitation. Because cloud isolation and tracking to examine the life stage are not necessary, the models are easy for applications and can be used in other areas as well provided the sampling time and meteorological conditions are similar.