The AC contactor plays a vital role as a low-voltage control device in electrical control systems. Accurate prediction of its electrical lifespan is of significant importance for ensuring system reliability and reducing maintenance costs. In this study, electrical lifespan experiments were conducted to collect degradation data of AC contactors. Time-domain features, including arcing duration, arcing energy, and average arcing power, were extracted from the arcing interval. In addition, statistical features such as root mean square (RMS), standard deviation, skewness, and waveform factor were computed to characterize the arcing behavior. Feature selection was then performed using Spearman correlation coefficient and mutual information to identify the most relevant indicators for lifespan prediction. The selected features were used as inputs to BiLSTM and CNN-BiLSTM. Both models demonstrated high prediction accuracy, with the CNN-BiLSTM model achieving superior performance in fitting the target labels. The results indicate that CNN-BiLSTM offers enhanced accuracy in predicting the electrical lifespan of AC contactors.