California Bearing Ratio (CBR) is an important indicator to evaluate base course materials and subgrade soils in pavement systems. This implementation of the machine learning (ML) approach in this research predicts CBR (California bearing ratio) values of the soil based on seven predictors for which values such as maximum dry density, soil classification, optimum moisture content, liquid limit, plastic limit, plastic index and swell can be easily obtained from the laboratory data using random forest (RF), decision tree (DT), linear regression(LR) and artificial neural network(ANN) models. Three hundred fifty-two soil samples that composed an experimental data set were classified in accordance with AASHTO M 145. They were divided into test data (20%) and training data (80%) in this model study. The performance of the models was evaluated with respect to standard statistical metrics like MSE (mean squared error), MAE (mean absolute error), RMSE, coefficient of determination and correlations. Based on these evaluation metrics, the random forest algorithm receives a smaller error and relative fashioned error (R2) values against other algorithms Thus a random forest algorithm developed using the analysis can make an accurate prediction of the soil’s CBR.