Currently the world marine primary production estimates range by a factor of two between different models. When the Arctic ocean alone is considered the factor rises to as much as fifty, because sea ice complicates remote sensing, and bio-optical properties of water as well as vertical distribution of phytoplankton differ from those of global waters. Arctic phytoplankton today deserve special attention as they are already living in waters with most prominent climate change effect, which are shifting towards fresher surface layer, thinner sea ice, more open water area and are very likely to experience ice-free summers in the near future. These shifts in turn alter solar irradiation, nutrient transport and plankton seasonality, and whether such an impact will result in an increase or a decrease of phytoplankton remains questionable. Since polar regions are difficult to access with the research vessels, field data are scarce here and remote sensing data provide an alternative. However it is not recommended to use remote sensing data alone, as the satellite ocean color algorithms are known to perform poorly at polar latitudes even if developed explicitly for the Arctic waters. Gaps in satellite data, which occur at these latitudes because of the presence of sea ice, clouds and low sun elevation angles, are also a source of error. The current study combined remote sensing, simulated and field data for the years 1998-2012 to investigate the seasonal cycle, variability and productivity of phytoplankton in the Greenland Sea, which is one of the most productive and field data-abundant regions of the Arctic. Specific objectives of our case Greenland Sea study were: 1) to study the interaction between phytoplankton and the physical factors, such as sea ice concentration and thickness, water temperature and salinity; 2) to investigate temporal trends, seasonal cycle and spatial variability of phytoplankton; 3) to obtain more accurate estimates of primary production. First the focus was set to northern part of Greenland Sea only (Fram Strait). The western, sea ice-dominated, part of Fram Strait proved to have short, late (middle May) and time varying phytoplankton blooms. At the marginal sea ice zone of the western Fram Strait a stratification induced by sea-ice melt was shown to have promoted phytoplankton growth, which resulted in the enhanced biomass observed in May. The eastern part, dominated by warm and salty Atlantic waters, experienced earlier (end of April) and longer blooms. In the eastern Fram Strait, stratification due to solar warming proved to act as a guiding factor for the the open ocean phytoplankton blooms, while the declining shelf ice was seen to promote the phytoplankton blooms along the coast of Svalbard. Secondly, we've proved that Greenland Sea chlorophyll vertical profiles largely deviate from the generalized schemes which were proposed for global waters and are usually used in primary production models. We then developed an Greenland Sea specific algorithm which allows accurate (4% underestimation when compared to in-situ data) estimation of chlorophyll vertical profile based on the surface value. This algorithm we then used to adapt global Antoine and Morel (1996) primary production model to Greenland Sea environment. When compared to field data, our Greenland Sea-adapted version of Antoine and Morel (1996) primary production model reproduced field data with better accuracy and was less biased (overestimation of in-situ data by 190 mgC/m2/day on an average) than the global Antoine model or the more commonly used satellite primary production model, Vertically Generalized Primary Production Model. According to our Greenland Sea-adapted version of Antoine and Morel (1996) model, annual Greenland Sea primary production was estimated as 161-194 TgC/yr, which is slightly higher than previously reported estimates. Monthly primary production values increased in the north-eastern open ocean area of the Greenland Sea, on about 120 mgC/m2 per day for the thirteen years of observations. The Greenland Sea primary production estimates we derived can be used for validation of primary production obtained from biogeochemical models, which are based on the climatology of nutrients. On local scale, it would be interesting for marine biologists to thoroughly compare our estimates with the benthic data to study the part of the carbon cycle which links increasing primary production of the Greenland Sea and the carbon stored in sediments.