Using datasets from one crypto exchange company in Indonesia, this study aims to predict churn in cryptocurrency exchange and analyze the factor that impacts it. The model developed in this work uses decision tree and random forest. Due to imbalance data, the undersampling method is applied before modelling. As the result, decision tree has 74.59% accuracy and the random forest 74.34%. From the two models, it is also found that the use of Google Authenticator and the frequency of cryptocurrency transactions are important factors to determine whether a customer will experience churn.