Abstract: To effectively predict the electric power energy consumption of iron and steel enterprises and improve the sustainability of resource utilization, an electric power energy consumption prediction model based on CNN-BiLSTM-Attention model is proposed. Firstly, the electric power energy consumption features are extracted by the mutual information (MI) method to optimize the input variables of the model and improve the generalization and robustness of the model. Then, a convolutional neural network (CNN) is introduced to further obtain the spatial correlation characteristics of multi-feature inputs, and a bidirectional long and short-term memory network (BiLSTM) is used to extract the sequence temporal change features, and finally, the extraction of key information is strengthened by the Attention mechanism (Attention). Comparison experiments are conducted with the actual electric power energy consumption data of an iron and steel enterprise, and the results show that the model has an average absolute error (MAE) and root mean square error (RMSE) of 3.9840 and 7.6907 on the dataset, respectively, which is superior to the other comparative models in terms of prediction accuracy. The results of the study have certain guiding significance for energy management and energy saving and emission reduction in iron and steel enterprises.