Abstract: To address the challenges in predicting superimposed thin sandstones in lacustrine basins,this study proposed a technical workflow for their prediction using the seismic meme inversion(SMI) method.First,an initial inversion model was constructed using log curves from sample wells selected based on seismic waveform similarity.Second,the initial inversion model was iteratively optimized in a Bayesian framework to yield high-resolution SMI results.Third,the SMI results were integrated with low-frequency inversion results to yield high vertical resolution while effectively characterizing the lateral superimposed patterns and boundaries of sand bodies.Case studies demonstrate that the SMI method achieved a prediction accuracy of 2 m to 3 m for superimposed thin sandstones.The coincidence rates for calibration and validation wells reached 91.5 % and 85.2 %,respectively,confirming the effectiveness of the SMI method and the high precision of the inversion results.Overall,this study provides an effective technical approach to predicting superimposed thin sandstones in lacustrine basins.