Abstract:California’s regulations on greenhouse gas emissions have led to increased adoption of Battery Electric Vehicles (BEVs) in the state. However, range anxiety remains a challenge for BEV drivers. To address this issue, accurately predicting BEV charging behavior and informing drivers when they need to charge is crucial. This study aims to develop a prediction framework based on a Bidirectional Long Short-Term Memory Network (BLSTM) model to suggest when BEV drivers should charge their vehicles. The BLSTM model will be trained using a robust dataset of trips and charges from a subset of the eVMT dataset collected between 2015 and 2020. By providing essential information, the BLSTM model can adapt to changes in driving and charging patterns. This study aims to provide a more accurate and precise method of predicting charging events than conventional machine learning models. Our results show that implementing our model could reduce unnecessary charging by 15-22%.