Carbon sequestration in agricultural soils under conventional land use that include tillage, fertilizing and removal of crops production, is to a great extent determined by accumulation of roots in soil after harvesting and their ability to be mineralized or to be incorporated into the stable organic compounds. Fertilizing can change substantially productivity of agroecosystems, and, therefore, plant residues accumulation in soil, and can reduce or stimulate stabilization of incoming carbon. As it is hard to estimate real stable carbon accumulation using laboratory and short-term field experiments results, we attempted to assess the possibility of carbon sequestration in arable soil in the in situ conditions on the basis of data of long-term (25 years) field experiment with fertilizers. The experiment was conducted on the soddy-podzolic sandy soil of Novozybkov Agricultural Experimental Station, Russia. The soil was very poor in plant nutrients. The initial soil (soil before treatments) had following characteristics: pÍCaCl2 – 4.5, organic carbon – 0.65%, total N – 500, mineral N – 10, exchangeable K – 42, available P – 100 mg kg-1 soil. Crop rotation, typical for the zone, included potato, lupine-rape mixture, oats and winter rye. Ammonium nitrate, triple superphosphate, potassium chloride, cow manure on the straw litter and dolomite lime were applied. Organic carbon content was estimated according to Turin technique. Carbon sequestration was calculated as a growth of organic carbon content in soil after 25 years of plants cultivation in comparison with organic carbon content in initial soil, %. The input of carbon into the soil depended mainly on plant residues accumulation, and, to a lesser degree, on manure application. The strong positive correlation between crop yields and carbon accumulation in soil was established (r=0.941). Sequestration of incoming carbon in soil organic matter was controlled by internal soil conditions favoring or limiting immobilization. The share of carbon of plant residues that was incorporated in organic carbon consisted of 13% in the unfertilized control, on the variant with manure – 19%, on the variant with NPK – 18%, on the variant with manure + NPK – 32%. Cultivation of plants during 25 years without fertilizers resulted in additional 8% of organic carbon content in soil; application of manure, NPK, manure + NPK favored 17, 26 and 54% of carbon sequestration in soil organic matter, correspondingly. Organic carbon content in the unfertilized soil increased on average of 25 years by 60, on variant with manure – by 132, on variant with NPK – by 204, on variant with manure + NPK – by 420 kg ha-1 year-1. Joint application of organic and mineral fertilizers favored not only the largest crop yields but as well much more effective sequestration of incoming carbon in soil organic matter compounds. Analysis of the long-term field experiment data leads us to the conclusion that fertilizing favored the real sink and accumulation of carbon in the studied extremely poor arable soil. General opinion that agricultural soils are CO2 sources is in all probability founded on the results of experiments carried out on more fertile soils where organic carbon decomposition exceeds its accumulation under disturbing factors such as tillage and fertilizing. Usage of the data of long-term field experiments can be a good tool for the verification of results of laboratory and short-term field experiments carried out on soils of different fertility.