Abstract: This paper proposes a Random Forest (RF) machine learning algorithm-based
prediction model for the state of charge (SoC) level of lithium-ion batteries for electric vehicles. To show the effectiveness of the proposed prediction model performance, the RF model has been compared with the other machine learning algorithms such as Support Vector Machines (SVM) and Gradient Boosting (GB) approaches. The dataset includes cell temperature, state of charge (SoC), voltage, and current readings at three different external temperatures—15, 25, and 30 degrees Celsius are considered in this paper to test the performances of the proposed model. After preprocessing of the dataset, 20% of the data was used for testing and the remaining 80% for training purposes. The various metrics such as mean squared error (MSE), mean absolute error (MAE), coefficient of determination (R^2), root mean squared error (RMSE), normalized root mean squared error (NRMSE), residual standard error (RSE), and relative absolute error (RAE) are usually preferred to evaluate the performance of the prediction models. The simulation results of the proposed model clearly show the effectiveness of SoC-level estimation for real-time battery management systems (BMS) when compared to other machine learning algorithms. The efficiency of the proposed model is 99% and execution time is less than 5 seconds. The accurate estimation of the SOC of lithium-ion batteries is crucial for optimizing battery performance, ensuring safety, and extending battery life in electric vehicles.