Abstract: The pretreatments for the measurement of dissolved inorganic carbon (DIC) isotopes in field water samples were developed from traditional BaCl₂ precipitation methods to directly continuous flow (GasBench-IRMS) methods. Further research needs be conducted to distinguish the difference between the traditional BaCl₂ precipitation methods and the continuous flow (GasBench-IRMS) method in order to select the best pretreatment for DIC. This study compared the results obtained by three pretreatment methods: BaCl₂ precipitation, medical sterile high-density polyethylene bottles and acidification of the sample in GasBench headspace vials in field work. Drip water and underground water samples were collected from Panlong cave in Guilin. The results show that because of CO₂ escape from the water sample, by using the BaCl₂ precipitation the value of the carbon isotope was more positive than other sample-pretreatment methods. The maximum deviation of the carbon isotope value from the underground river is 0.26. The maximum deviation of the carbon isotope value from the cave drip water is 0.33. The temperature and atmospheric pressure are similar in the field and the laboratory, therefore it did not cause solubility change of CO₂ and HCO^-₃ in water samples. The DIC results of the medical sterile high-density polyethylene bottles and GasBench headspace vials are the same. The GasBench headspace vials pretreatment method can effectively avoid the solubility change of CO₂ and HCO^-₃ caused by environmental changes, which leads to the CO₂ escape from the water sample or dissolved into the water sample from the atmosphere of CO₂. This change may cause carbon isotopic fractionation in DIC of water samples. Using GasBench headspace vials, which directly produce and collect CO₂ gas in the field is the recommended DIC pretreatment method.