Abstract: Instruments of elemental analysis-isotope ratio mass spectrometer (EA-IRMS) and GasBench-isotope ratio mass spectrometer (GasBench-IRMS) were used in the determination of ratios of stable isotopes of carbon, nitroge, hydrogen and oxygen (abbrev. as δ¹³C、δ¹⁵N、δ²H and δ¹⁸O respectively) in 83 milk samples collected from 7 different countries of the world (namely. Australia, New Zealand, Spain, Germany, Austria, Italy and China). And by using the software of SPSS 20.0, the data of δ of the 4 elements obtained were sudied chemometrically by variance analysis, clustering analysis and discriminate analysis for the producing area of the milk samples. It was found from the testing results that some valuable informations relating to the producing source of the milk samples were obtained from each of the δ values of the 4 elements, however it was still not a sound base for discrimination of the producing area of the milk samples. It was shown by the results of further investigation, the correctness of discrimination was significantly raised to a value of 84.3% by a combinatorial study of the data of δ¹³C、δ¹⁵N and δ²H, and discrimination of the producing area of milk samples could be attained more effectively. Furthermore, mathematical models for discrimination of producing areas among the 7 countries were also established on the base of the δ-values of the 3 isotopic elements. In discrimination of unknown samples from the 7 countries, the isotopic ratios of carbon, nitrogen and hydrogen were determined and the data obtained were substituted into the discrimination models to give values of Y, among which the one with largest Y value was qualified belonging to the country designated by the model.