Abstract-In this paper we analyze the performances of a new probabilistic belief transformation, denoted DSmP, for the sequential estimation of target ID from classifier outputs in the Target Type Tracking problem (TTT). We complicate here a bit the TTT problem by considering three types of targets (Interceptor, Fighter and Cargo) and show through Monte-Carlo simulations the advantages of DSmP over the classical pignistic transformation which is classically used for decision-making under uncertainty when dealing with belief assignments. Based on our previous works for the justification of rules of combination for TTT problem, only the Proportional Conflict Redistribution rule and the hybrid fusion rules are considered in this work for their ability to deal consistently with high conflicting sources of evidence with three different belief assignment modelings.