Preventing Overfitting in Tree Models

Imagine you're developing a classification model. What strategies would you employ to prevent overfitting when working with tree-based algorithms?