This paper presents the novel and efficient warhead tracking filter based on probabilistic data association (PDA) scheme for FMCW radar seeker with anti-ballistic capability. In frequency domain, the warhead tracking problem can be cast into a difficult data-association problem because multiple and closely located frequency measurements are generated by scatterers of a ballistic missile and multi-path clutters. For enhancing the warhead detection and tracking performance, the proposed warhead tracking filter adopts the feature information about scatterers' appearance. The feature information indicates the possibility that the measured frequency measurement is originated from the scatterer in warhead section of ballistic missile. Utilizing the feature information, the proposed tracking filter is able to accurately separate the warhead frequency measurement from others. Moreover, for its simple structure, it is suitable for real-time implementation. Simulation results demonstrate the effectiveness and the reliability of the proposed solution over the conventional PDA based filter.