During the stage separation phase of ballistic missile (BM), in addition to the warhead, dierent types of confusing objects such as fuel tanks and boards occur. At this point, classifying warhead and other confusing objects is very important for missile defense. Therefore, we propose tracking and classication algorithm in the BM stage separation environment through Rao-Blackwellized particle lter. The proposed algorithm utilizes the known physical character-
istics, aerodynamic forces, moments of the targets in advance for dynamic update and RCS information for measurement update. The performance of the proposed framework is analyzed by several simulation results.