Abstract: In China, lacustrine carbonate rocks are widely distributed and show great prospects for petroleum exploration. The high-content terrigenous debris and diversified lithofacies lead to great difficulty in identifying lithofacies and severely restrict the exploration expansion and efficient development in this field, which is typically represented by the upper part of the Paleogene Lower Ganchaigou Formation (E₃²) in Yingxi area in Qaidam Basin with a burial depth of greater than 4000 m. Therefore, a large number of core data, thin section and experimental analysis data in Yingxi area are studied in detail, and lithofacies and logging response characteristics of lacustrine carbonate rocks are systematically analyzed to establish the lithofacies identification method and chart. The study results show that five types of lithofacies are identified in E₃² in Yingxi area, including granular lime-dolostone, micrite lime-dolostone, laminated lime-dolostone, argillaceous gypsum and lime-dolomitic mudstone. The analysis of logging response characteristics indicates that GR, matrix density and bulk density are sensitive parameters for lithofacies identification, and rock fabric factor ( RFF ) technology is proposed for the first time. The calculated RFF and element logging data are integrated to establish the lithofacies identification chart, which is applied to 48 wells in the study area. The average coincidence rate of lithofacies interpretation reaches up to 80.2% in 12 coring wells, showing good application results. RFF technology supports to improve the lithofacies identification accuracy of lacustrine carbonate rocks, and is of great significance to determine the reservoir distribution law and guide the petroleum exploration and development in practice.