The paper addresses a crucial problem of accurate range estimation and state of charge (SoC) determination in electric vehicles (EVs) to improve charging as well as discharging decision-making. Reliable real-time range prediction and precise SoC estimation remain significant research challenges in EV technology. A multivariate linear regression (MLR) model was developed and trained on a dataset of 195 EVs, capturing key attributes including Power_Train, and Body_Style, top speed, acceleration, efficiency, Fast_Charge, Rapid_Charge. Comprehensive data cleaning and feature engineering were performed using Pandas and scikit-learn to enhance model accuracy. Existing models like decision tree regression and random forest regression suffer from overfitting and static decision thresholds. The study addresses these limitations by integrating MLR with rule-based SoC decision-making for real-time EV range predictions. The model’s efficacy is demonstrated by the experimental results, which show low mean abs error is 8.96 and root MSE is 11.73, along with high R2 scores 0.9835 with training data and 0.9807 with test data. The outcomes suggest that the model may be able to provide precise range and SoC predictions, enabling EV users to make better-informed decisions about charging and discharging.