To recognize the lithology of strata is a very important task of such geological studies as stratigraphic correlation and sedimentary facies analysis. When there are scanty core data and more well log data, multivariate statistical analysis is a very effective method to recognize the lithology of strata. So, on the basis of the data from the Yong-1 and Sha-4 conglomerate body of the Shengli oilfield and through an analysis of the correlative characters of drilling cores and well logs, some well log parameters that are usable in recognizing lithology have been selected by using discriminative analysis and the related functions of lithological recognition has been determined. By using this function the lithologies of strata at different depth points can be reflected very quickly, and the lithological column can be drawn, which provides the basic data for a further geological study.