The ESA EOStat and SATMIROL projects exploit the new capabilities of satellite data from the Copernicus programme, aiming at creation of a system of support for acquisition of statistical information for agricultural production in Poland. We demonstrate and validate the machine-learning-based approaches for a country-scale field-level crop type mapping and crop yield forecasting at NUTS-2 and LAU-2. The main user of developed data products is Statistics Poland currently exploring innovative data sources based on satellite Earth Observation to support official statistics.