Recent studies have proposed barium isotopes as a novel proxy for studying primary productivity in paleo-oceangraphical studies and elements cycling through the critical zone. Pristine marine carbonates are generally assumed to preserve Ba isotope compositions of ancient seawater. However, Ba incorporated in or adsorbed on detrital minerals such as clays in impure carbonates may limit the accurate application of the Ba isotope proxy for paleo-ocean environmental reconstruction purposes. We present here a sequential extraction procedure and show that a considerable range of Ba concentrations can be associated with the four operationally defined sequential leaching fractions (water-soluble, exchangeable, carbonate, and oxidizable fractions). Chemical separation of Ba from these leachates is achieved with a recovery of >98.6% by our modified ion exchange procedure. Potential instrumental mass bias effects and barium isotope fractionation during the chemical separation are corrected using a carefully optimized 130Ba-135Ba double-spike method. A long-term reproducibility better than ±0.03‰ (2SD) for δ137/134Ba has been achieved by using the double spike-thermal ionization mass spectrometry (DS-TIMS) in this study. We demonstrate that significant variations of δ137/134Ba in the analyzed leachates suggest a considerable Ba isotope fractionation between carbonate mineral phase and noncarbonate phases of marine carbonate rocks. The barium isotope distribution in a set of standard reference materials and natural geological samples under various geological settings has been presented. When utilizing Ba isotopes as a proxy for primary productivity and the biogeochemical cycling of Ba, our new findings from sequential Ba extraction as well as our modified precise DS-TIMS analytical setup should be taken into account.