Molodtsov [35] initiated the soft set theory, which can be used as a generic mathematical tool for dealing with uncertainty. Maji [30] defined the concept of neutrosophic soft set which is based on a combination of the neutrosophic set and soft set models. In this paper, we give the notion of neutrosophic soft set with a new style and present some algebraic properties. We define several distance measures between neutrosophic soft sets and give an axiomatic definition of neutrosophic entropy for a neutrosophic soft set. To find the most ideal alternatives from all possible alternatives, we propose a method based on similarity measure. Thus we can rank all possible alternatives. Finally, the proposed similarity measure is applied to a multicriteria decision making problem.