Although deep learning is dominant in diverse application fields, decision trees and their ensembles do not vanish. Bayesian decision trees introduce the Bayesian thoughts into the tree-based models. In another word, they involve some piror probability distribution when constructing the decision trees.