Bad credit card is a problem of inability of credit card users to pay credit card bills that can cause losses to both parties concerned. In order to avoid losses caused by bad credit cards, the provider must conduct a careful analysis of prospective or old customers using credit cards. This study aims to classify bad credit card customers using machine learning techniques, namely classification techniques. One of the classification techniques used is the XGBoost method which is useful for regression analysis and classification based on the Gradient Boosting Decision Tree (GBDT), the XGBoost method has several hyperparameters that can be configured to improve the performance of the model. Hyperparameter tuning method used is grid search cross validation which is then validated using 10-Fold Cross Validation. XGBoost hyperparameters configured include n_estimators, max_depth, subsample, gamma, colsample_bylevel, min_child_weight and learning_rate. Based on the results of this study proves that the use of algorithms with hyperparameter tuning can improve the performance of eXtreme Gradient Boosting algorithm in the process of classification of credit card customers with an accuracy of 80.039%, precision of 81.338% and a recall value of 96.854%.
 
 Keywords: XGBoost, classification, Accuracy, Precision, Recall