Decision trees are widely known models in Supervised Machine Learning with efficient inference mechanisms and outstanding interpretability. In this article, we design the implementation of classical Inductive Decision Trees under a quantum computing paradigm, and explore the advantages of Quantum Decision Trees designed in the presence of missing and uncertain data. Our findings extend to quantum ensembles analogous to Decision Forests as a Quantum Machine Learning method to improve the interpretability of a type of variational quantum circuits. Our approach provides an improvement in efficiency in the case of probabilistic inference with respect to the classical counterpart, and a general methodology is designed to address multiple classification tasks with Quantum Machine Learning tools, with a focus on the interpretability of quantum models. The theoretical results are supported by experimental simulations using different data sets and state-of-the-art examples.