The Baltic Sea is annually covered by sea ice, which is structurally similar to the sea ice in polar oceans. Structural similarity to polar sea ice decreases northwards with decreasing salinity of parent water. The Baltic Sea is heavily influenced by riverine inflows and allochtonous dissolved organic matter (DOM). One of the climate change scenarios is increase in precipitation which will most probably increase also allochtonous DOM load in the Baltic Sea. Biogeochemical cycling of nutrients and DOM takes place within the ice matrix by bacteria. They are a central heterotrophic organism group in sea-ice ecosystems, especially in carbon and nutrient transformation and transfer. Bacteria incorporate DOM and thus transfer it to upper trophic levels. They also actively participate in nutrient transformations between particulate, dissolved and gaseous forms in ice. These processes can have large-scale consequences for wintertime nutrient and carbon budgets that are up to date insufficiently understood. Especially DOM and nitrogen dynamics in sea ice are different than in open-water environments. Changes in ice physical environment or quality of DOM can alter bacterial community composition and successional dynamics that in turn affect sea-ice biogeochemistry. To be able to assess the role of sea ice in current, and predict the future, climate regimes, better understanding of sea-ice bacteria and their biogeochemical functions is clearly needed. We conducted a sea-ice mesocosm experiment with three different types of organic substrates in order to study the effects of substrate alteration on bacterial community physiology and composition. Diatom- and dinoflagellate-derived DOM as well as sucrose were added to three of the tanks and compared with a control tank without any additions. The experiment was done in four 450 L tanks in the cold laboratory of the Finnish Institute of Marine Research (presently Finnish Environment Institute). The tanks were filled with natural Baltic Sea water retrieved from the Gulf of Finland shortly before the experiment started. Ice and water samples were collected from the tanks weekly for 3 months in February–May 2007. The effects of the added substrates were assessed by measuring bacterial production and biomass increase, bacterial community composition and nitrogen fluxes including DON. Bacterial community compositions were analysed with cloning and 16S rDNA terminal restriction fragment length polymorphism (T-RFLP). Results show that bacterial activity increased resulting from DOM additions in ice and water phases. Clear differences in bacterial community composition between treatments were also detected.