Abstract: Oil and gas exploration degree of the Shanan Sag of the Bohai Basin is relatively low due to lack of drilling cores and source rock samples. TOC logging prediction models for source rocks are established in this paper for the Members of E₃d₃、E₃s₁₊₂、E₂s₃ by the multiple linear regression method and the improved ΔlgR method based on the logging data and limited measured TOC data of the mudstone samples collected from six single wells. After comparison and optimization, it is found that both the improved ΔlgR method and the four- parameter linear regression method are effective to predict the TOC vertical distribution pattern of the lacustrine source rocks in single wells in the Sag and the improved ΔlgR method is obviously more efficient. Therefore, predicted TOC curve of improved ΔlgR is selected to evaluate the source rocks for two representative single wells. The result shows that the E₂s₃ is characterized by high ratio of source rocks, high TOC content and large thickness of high-quality source rocks. It is the main source rock of the study area and the Shanan sag must have good prospect for oil and gas exploration. The research results are helpful to the evaluation of source rocks in the Shanan sag and will provide solid scientific basis for further understanding of oil and gas exploration potential in the study area.