Abstract: The Support Vector Machine (SVM) Method is first introduced into reservoir fluid property identification, by which the models for identifying oil, gas and water are developed through study of the complex relations between measured logging data and comprehensively interpreted parameters about reservoir fluids.The applied results indicate that SVM method is a feasible, effective and higher accurate way for well logging interpretation of reservoir fluids as a new, simple and reliable method with high accuracy.

Key words: support vector machine, fluid, identification, well logging interpretation