Nowadays the dominant oil and gas exploration theory has gradually transformed from “Source control theory”, “Belt controlled theory” to “Facies control theory”, and the exploration target mainly aims at the subtle reservoirs such as the lithologic and stratigraphic reservoirs, the reservoir petrophysical facies shows great prospects and potential in predicting high quality reservoirs, evaluating reservoir parameters and researching the distribution of residual oil and gas since its being proposed in 1990s. Though the division of petrophysical facies has always been a hot topic among petroleum geologists, there is still no common understanding on the division method of petrophysical facies at present. The division methods of petrophysical facies can be mainly classified into two types: one is mathematical method; the other is dividing the petrophysical facies mainly through various geological data and related studies. But the applicable conditions of these two types of division methods are limited, and there are defects and disadvantages existing in the division schemes, what is worse, the naming of the divided petrophysical is not scientific. This article, based on a large amount of literatures at home and abroad, proposes a division method of petrophysical facies under the guidance of “facies control theories” according to the present situation and problem, that is, dividing the petrophysical facies through the superiposition and coupling of the main control factors of petrophysical facies: sedimentary facies, tectonic facies and diagenetic facies. The second Member of Upper Triassic Xujiahe Formation in Penglai Area is taken as an example to illustrate and testify the feasibility and accuracy of this method. The petrophysical facies of Xu 2 reservoir is divided and named by this method, then the quantitatively characterization method is proposed, and then the relationships between petrophysical facies types and oil and gas productivity are discussed and analyzed. Thus the division results of petrophysical facies can be used to predict and optimize the high quality reservoir bodies. The proposed method on the one hand can avoid the errors caused by the previous methods since it divides petrophysical facies from the aspects of geological facies, and the geological factors affecting the petroleum accumulation are fully taken into account, on he other hand, it can provide a more scientific naming scheme of petrophysical facies.