Colored dissolved organic matter (CDOM) widely exists in water, accounting for 10%-90% of dissolved organic carbon (DOC). In order to analyze the distributions of DOC concentrations in the reservoirs in the northeastern China, five field campaigns were carried out in Baishan Reservoir and Fengman Reservoir in the Songhua River Basin from 2015 to 2020. Based on the analysis of the absorption characteristics of CDOM coefficient (355 nm), the remote sensing inversion model of CDOM absorption coefficient at 355nm (R²=0.82) is established by band ratio method based on Landsat series satellites and the DOC concentrations of the reservoirs are estimated according to the strong correlations between CDOM and DOC. The results show that: The CDOM concentrations of inland reservoirs in the northeastern China can be well retrieved by using the in-situ sampling data and Landsat datasets; The annual spatial-temporal distributions of CDOM and DOC concentrations in the Baishan Reservoir and the Fengman Reservoir from 2000 to 2020 are with low fluctuations, and the concentration of CDOM and DOC appeared an upward trend after 2010. The Baishan Reservoir and the Fengman Reservoir have lower M values (the ratio of absorption coefficient of CDOM at 250 nm and 365 nm) and S_275-295 (the slope of the absorption spectrum of CDOM at the 275-295 nm), and higher SUVA₂₅₄ (the ratio of CDOM absorption coefficient at 254 nm to DOC concentration), indicating that the humic contents in the CDOM compositions are higher, and tend to be consisted of terrigenous macromolecules. Using multiple regression analysis method to analyze the annual driving factors of CDOM distributions of the reservoirs is mainly affected by rainfalls, fertilizer applications and sewage discharges. Finally, this study enriches the understanding of the variations of CDOM concentrations in typical reservoirs in the northeastern China and provides a reference for the construction of DOC inversion models and contributes to the accuracy estimation of DOC concentrations.