Electric submersible pump oil recovery is currently one of the most important oil recovery methods in offshore oil fields. Its failure can affect the normal production and operation of oil wells and cause economic losses. Therefore, it is particularly important to predict the remaining service life of electric submersible pumps and prevent failures. To ensure the high-quality operation of electric submersible pumps, a remaining service life prediction method based on ensemble learning model is proposed according to the data characteristics of electric submersible pumps. Firstly, the remaining service life at each time point is calculated as the label function, and the random forest algorithm is used to screen the high contribution feature parameters input model. An ensemble model consisting of SSA-CNN and SSA-LSTM base models weighted by absolute error is constructed. On site data verification shows that the two base model algorithms have their own advantages and disadvantages in different situations. SSA-CNN has more advantages during data fluctuation periods, while SSA-LSTM has more accurate overall predictions. When the same data is input into the ensemble model, it is found that the prediction error of the ensemble model is significantly smaller than that of the two base models, combining the advantages of both. There is a significant improvement in overall accuracy and evaluation accuracy during the change stage. The actual calculation verification shows that the prediction accuracy of the integrated model is 6.41% higher than that of the base model, which is significantly improved compared to existing methods and has strong robustness and stability.