The accuracy of sedimentary facies and facies interface identification is not high due to the complex grain size composition and low core recovery in sandy sediment. In this paper， a lithology identification model based on the support vector machine （SVM） is established based on core and logging data， and the influence of the training set size on the identification accuracy of the model is analyzed and compared with the BP neural network model. The results show that the SVM model has high accuracy of lithology identification and low demand for training samples， which can effectively make up for the insufficient core recovery and reduce the drilling cost. It is feasible to use the logging lithology identification model based on SVM to identify the sedimentary sequence automatically in unconsolidated sediment investigation， which is a beneficial attempt to implement green exploration.