A new method is proposed which allows for reasonably accurate cloud classification based upon Gaussian approximation of cloud emission and using a pair of visible-infrared, high resolution METEOSAT images. Its most important characteristic is the very low amount of CPU time required for a single classification. It becomes then suitable for application in very short range weather forecasting (nowcasting) by ensuring the adequate time coverage and necessary rapidity of use for the operational environment. Two classifications of summer and winter meteorological situations are presented.