Colored dissolved organic matter (CDOM) plays a crucial role in aquatic ecosystems, influencing the cycling of nutrients and the transport of pollutants in water bodies. This study focused on the remote sensing inversion of CDOM in the Furong River Basin, a karst area, utilizing in-situ data and Sentinel-2A remote sensing imagery. The CDOM inversion model was constructed using AdaBoost and Local Interpretable Model-agnostic Explanations (LIME) algorithms. The results exhibited that the model has high inversion accuracy, with an R2 of 0.86 and an RMSE of 0.495 /m for the test-set. There was spatial variability in CDOM between the dry and wet seasons, with a higher absorption coefficient generally observed during the wet season. Additionally, interpretable technology revealed the contribution differences of various band ratios to the model inversion, with band ratios B4/B3, B6/B4, and B5/B2 having a substantial impact. The study also found that different driving factors affect CDOM in different seasons, higher temperatures and lower precipitation tend to reduce CDOM, land use significantly impacted river CDOM, and the negative correlation between population density and CDOM may be related to the dilution effect of the regional groundwater system and environmental protection measures.