Meteosat Second Generation is a geostationary satellite, designed to explore atmospherical processes. Main instrument on board MSG is SEVIRI, a 12-channel scanning radiometer, observing the Earth full disk in nominal MSG position at 0° longitude with a horizontal resolution of 3 km at nadir. In this paper one of MSG Cloud Products – Cloud Mask (CMa) was used to analyze cloud detection over Poland. The main objective of Cloud Mask is to discriminate all cloudfree pixels, because correct cloud detection is an important pre-processing step, to use many different MSG products. As the result of this algorithm all pixels are categorized in five categories as cloudy, cloud contaminated, cloud-free, snow/ ice fi lled or no processing. The aim of this study is to compare Cloud Mask data to a standard synoptical observations. Ground observations classified cloudiness in a 9-degree scale (octas) and normalization this different scales was the greatest methodical problem. Base sources was 12 situations cloud cover, represented varied types of cloudiness over Poland, in effect 812 sat-synop observations. This satellite data was compared with ground observations through contingency tables and statistical indicators. The percentage of correct observations for 5x5 pixel matrix equals 27,3. As the results shows, it is visible linear relationship between satellite and synop data with correlation coefficient equals 0.73. Most important effect of this study is to indicate a underestimating satellitebased cloudiness observations. The mean error (satellite-synop) ranging between +0,6 and –1,3 octas, and except cloud free situations, is negative.