Abstract: During the last years, it has been demonstrated that δ¹⁸O isotope analysis of ethanol in juices and alcoholic beverages provides powerful information to assess their adulteration by addition of water, or to authenticate their geographical origin. However, it is difficult to accurately and conveniently quantify it due to the measurement principle and its water miscibility. To eliminate the impact of water on ethanol δ¹⁸O analysis, a porous-polymer-bonded GC column was used to achieve a baseline separation of ethanol and water, then the water was vented by backflush function and the ethanol was converted and analyzed. This is the first manuscript presenting a systematic evaluation and characterization of a method for δ¹⁸O isotope analysis of ethanol in aqueous samples by direct injection of the sample (diluted with acetone) into a GC-TC-IRMS system, with the consequent benefits of eliminating sample pre-treatment, minimum sample amounts required and minimum sample throughput. This method has the following advantages: only 70-200 μL of the sample was required for up to 300 possible injections, and high throughput was observed, caused by short analysis time (18 min for each run), the influences of oxygen-containing compounds on ethanol δ¹⁸O analysis were eliminated by the use of a capillary-column bonded porous polymer to obtain a baseline separation prior to high-temperature conversion. Precision was determined to be less than 0.5‰ (1σ), and accuracy was validated by spiked samples and proficiency test samples.