This paper describes the application of a Differential Evolution based approach for inducing oblique decision trees in a global search strategy. By using both the number of attributes and the number of class labels in a dataset, this approach determines the size of the real-valued vector utilized for encoding the set of hyperplanes used as test conditions in the internal nodes of an oblique decision tree. Also a scheme of three steps to map the linear representation of candidate solutions into feasible oblique decision trees is described. Experimental results obtained show that this approach induces more accurate classifiers than those produced by other proposed induction methods.