Cuttings logging is a direct identification of stratigraphic lithology and oil content. The proper description of lithology plays an important role in cuttings logging. The laser-induced breakdown spectroscopy (LIBS) of seven elements including Mg, Si, Al, Fe, Ca, Na and K are selected as the analysis lines, and the four kinds of cuttings lithology of marlite, mudstone, shale and sandstone are identified combining with three chemometric methods including principal component analysis (PCA), soft independent modeling of class analogy (SIMCA), supervised Kohonen networks (SKNs). The average correct recognition rates of SKNs and SIMCA models are 93.75% and 78.75%, respectively. Results show that the combination of LIBS technology, PCA and nonlinear SKNs methods can realize the effective lithology identification of cuttings having similar physical properties and chemical composition.