The regression Tree is about a predictive model in decision making, an SVM model is CLASSIFICATION as representation of the point examples in space, mapped so that the examples of separate categories are divided by as wide a clear gap as possible."good separation", Now the use of each method depent on your need ,or rather your problematic.
Both "Support Vector Regression" and "Regression Tree" are able to handle a regression problem. In contrast, SVM and DT (for classification) are dealing with a classification task. The performance of those two algorithms, in each learning type, depends on their parameters as well as the characteristics of the data.