Abstract: A system and method for predicting vehicle battery failures includes maintaining a rolling buffer of voltage data from a vehicle and analyzing the data during engine state transitions. The system detects transitions from engine off to on states and analyzes voltage data from a predetermined time period preceding the transition to identify minimum voltage values. These values are compared against configurable threshold voltages to assess battery health. The system generates time series models from historical voltage data to predict future voltage values and potential battery failures. The system generates customized alerts through a graphical user interface indicating predicted time until battery failure and recommended maintenance actions.