Dissolved black carbon (DBC) plays a key role in global carbon cycle and pollutant transport. However, the time-consuming and labor-intensive chemical analysis limits its spatiotemporal resolution. Here, we developed models to predict DBC from chromophoric dissolved organic matter (CDOM) measurements across the land-to-ocean continuum. We found that the mean ratio of DBC to light absorbance at 254 nm (a254) changed <20% among different environments. However, a single-wavelength model is inadequate for precise prediction due to microbial production of CDOM. Incorporating longer wavelengths using multiple linear regression improves model performance. Random Forest Regression using the full spectral range performed even better at all environments, including the open ocean, achieving a root mean square logarithmic error of <0.15, median symmetric accuracy of <10%, and R2 of >0.85. This study demonstrates the feasibility of using CDOM to predict DBC concentrations and highlights the potential for in situ monitoring and remote sensing applications.