This article proposes a VMD-SSA-BiLSTM model for predicting the remaining useful life (RUL) of supercapacitors under different operating conditions. Supercapacitors are key energy storage devices with high energy and power density, wide temperature range, and long lifespan. Accurate RUL prediction is crucial for safe and efficient operation of energy storage systems. The model combines variational mode decomposition (VMD) to decompose aging data and remove noise, and a bidirectional long short-term memory (BiLSTM) neural network for prediction. The sparrow search algorithm (SSA) is used to optimize the hyperparameters of the BiLSTM network, improving the consistency between input data and network structure. The model was tested under various temperatures and voltages, showing high prediction accuracy and robustness. The results indicate that the VMD-SSA-BiLSTM model outperforms traditional methods, with an average RMSE of 0.112519 and a minimum of 0.031426. The model effectively handles noise and capacity recovery in degradation curves, providing accurate RUL predictions under different operating conditions. The study highlights the importance of accurate RUL prediction for maintaining the performance and reliability of supercapacitors in various applications.This article proposes a VMD-SSA-BiLSTM model for predicting the remaining useful life (RUL) of supercapacitors under different operating conditions. Supercapacitors are key energy storage devices with high energy and power density, wide temperature range, and long lifespan. Accurate RUL prediction is crucial for safe and efficient operation of energy storage systems. The model combines variational mode decomposition (VMD) to decompose aging data and remove noise, and a bidirectional long short-term memory (BiLSTM) neural network for prediction. The sparrow search algorithm (SSA) is used to optimize the hyperparameters of the BiLSTM network, improving the consistency between input data and network structure. The model was tested under various temperatures and voltages, showing high prediction accuracy and robustness. The results indicate that the VMD-SSA-BiLSTM model outperforms traditional methods, with an average RMSE of 0.112519 and a minimum of 0.031426. The model effectively handles noise and capacity recovery in degradation curves, providing accurate RUL predictions under different operating conditions. The study highlights the importance of accurate RUL prediction for maintaining the performance and reliability of supercapacitors in various applications.