Abstract <p>Dents are common defects in oil and gas pipelines. Accurately and efficiently predicting the strain in dented pipelines is crucial for their safety assessment. Currently, there is limited research on strain prediction models for dented pipelines both domestically and internationally, and related work on predicting the mechanical behavior of dented pipelines is relatively immature. In view of this, this study proposes a model for predicting the maximum equivalent plastic strain at dented areas of pipelines using a Back Propagation Neural Network (BPNN). In this study, a static analysis model for dented pipelines was established using finite element software, and its accuracy was verified through relevant experiments. Based on the finite element model and combined with the Pearson correlation coefficient method, the interdependencies between maximum equivalent plastic strain and various key parameters were analyzed. A comprehensive training dataset was obtained by ranking the parameter correlations. Using the dataset, a strain prediction model was established through the application of the backpropagation algorithm and optimization of the number of neurons in the BPNN. The model was used to predict the maximum equivalent plastic strain at pipeline dents, and the stability of the model’s predictions for the maximum equivalent plastic strain was validated against experimental data and random datasets. The results indicate that the prediction errors for experimental data and the random dataset are minimal, demonstrating that the model can accurately predict the strain behavior of dented pipelines. The predictive model established in this paper can serve as a reference for assessing pipeline dents in practical engineering.</p>