In the current field of log interpretation based on machine learning, due to the difference in data distribution, it is difficult to directly apply the model trained on the existing log data to the new log interpretation. This paper focuses on the intelligent interpretation of geophysical logging, and proposes a semi-supervised model fine-tuning method, Log2FT, with the help of machine learning. In this method, the model is trained on the source domain, and then fine-tuned with a few labels in the target domain to improve the adaptability of the model. In order to verify the effectiveness of this method, we selected four wells D, E, F and G located in Jiyang Depression, Bohai Bay Basin, and conducted four groups of experiments D→E, E→D, F→G, and G→F, respectively. Through a series of experimental designs, including parallel repeated experiments, contrast experiments, ablation experiments and related interpretation analysis, the effectiveness and practicability of the proposed method are fully verified. The experimental results show that this method significantly improves the accuracy of logging interpretation.Thisresearchhelpsto overcome the problem of data distribution difference in the existing logging interpretation, and provides a feasible and effective method for interpreting new logging data.