This paper presents a novel method for predicting the Remaining Useful Life (RUL) of supercapacitors under different operating conditions, integrating Variational Mode Decomposition (VMD) with a Bidirectional Long Short-Term Memory (BiLSTM) neural network. The study begins by conducting aging experiments on supercapacitors under various temperatures and voltages to obtain aging data. VMD is then applied to decompose the aging data, which helps eliminate disturbances such as capacity recovery and test errors. The hyperparameters of the BiLSTM are optimized using the Sparrow Search Algorithm (SSA) to improve the consistency between the input data and the network structure. The decomposed aging data are input into the BiLSTM for prediction. The experimental results show that the proposed VMD-SSA-BiLSTM model has high prediction accuracy and robustness under different temperatures and voltages, with an average RMSE of 0.112519, a decrease of 44.3% compared to the BiLSTM, and a minimum of 0.031426. The model's effectiveness is demonstrated through comparisons with other methods, and its ability to handle complex degradation patterns is highlighted. The study concludes that the VMD-SSA-BiLSTM model is a reliable tool for predicting the RUL of supercapacitors, offering significant advantages over traditional methods.This paper presents a novel method for predicting the Remaining Useful Life (RUL) of supercapacitors under different operating conditions, integrating Variational Mode Decomposition (VMD) with a Bidirectional Long Short-Term Memory (BiLSTM) neural network. The study begins by conducting aging experiments on supercapacitors under various temperatures and voltages to obtain aging data. VMD is then applied to decompose the aging data, which helps eliminate disturbances such as capacity recovery and test errors. The hyperparameters of the BiLSTM are optimized using the Sparrow Search Algorithm (SSA) to improve the consistency between the input data and the network structure. The decomposed aging data are input into the BiLSTM for prediction. The experimental results show that the proposed VMD-SSA-BiLSTM model has high prediction accuracy and robustness under different temperatures and voltages, with an average RMSE of 0.112519, a decrease of 44.3% compared to the BiLSTM, and a minimum of 0.031426. The model's effectiveness is demonstrated through comparisons with other methods, and its ability to handle complex degradation patterns is highlighted. The study concludes that the VMD-SSA-BiLSTM model is a reliable tool for predicting the RUL of supercapacitors, offering significant advantages over traditional methods.