Due to factors including high prices, limited supply, financial inability, and other factors, newly built cars are unable to reach buyers despite the significant expansion in car usage. As a result, the used automobile industry is growing rapidly all over the world, but it is still in its infancy in India and is largely dominated by the unorganised sector. This creates the potential for deception when purchasing a used car. Thus, a highly accurate model that can estimate the cost of a used car without favouring either the customer or the merchandiser is needed. This model develops an XGBoost Regression model based on supervised learning that can learn from the input automobile dataset. This project offers a low error workable model for estimating secondhand car prices. For dependable and accurate forecasts, a large number of unique attributes are considered. The findings achieved are in line with theoretical predictions, and they outperform models that rely on straightforward linear models. The XGBoost Regression algorithm is used to build a model, and for comparison, other machine learning algorithms such as Linear Regression, Random Forest Regression, Lasso Regression, Ridge Regression, Decision Tree Regression, ElasticNet Regression, Adaboost Regression, and Gradient Boosting Regressions are also built. The automobile dataset is used to test these techniques. According to experimental findings, the XGBoost Regression model has a training accuracy of 99.86%, a test accuracy of 94.45%, and a root mean square error of 0.05237.among all the other methods, has produced the least inaccuracy. The work reported here has significant ramifications for future research on the XGBoost Regression model for Used Car Price Prediction, and it may one day contribute to a 100% accurate solution to the fraud problem.