Abstract: For reducing the negative influence of chaff on the ture target’s track and attack of the terminal guidance radar seeker, a novel anti-chaff-jamming method based on geometrical inference matching algorithm is proposed. As the chaff cloud is made up with amount of dipoles, whose RCS, disperse equation and speed varies as the function of time, so it is unstable for the high resolution range profile (HRRP) of chaff cloud, which shows as the number, scattering amplitude and relative position of different scattering centers are time-varying; but the relative position of scattering center originated from targets stays constant when the incident beam varies within a small range. In the algrithm, all possible scattering centers extracted from two consecutive HRRPs after motion compensation are regarded as point patterns, then a geometrical inference matching algorithm that can work under partially matching problem with Euclid transformation, is used to form the optimum matching group which has the maximum number of consistent correlative points. The point patterns in the optimum matching group are the target scattering centers, therefore the geometrical center can be calculated from the scattering centers. In addition, the translation distance of the target between two consecutive HRRPs can be used to estimate the target velocity, which is used to compensate the next HRRP and forecast the target position of next time. For velocity measurement, the presented algorithm overcomes the limitation that the conventional algorithm can not work effectively when the chaff jamming exists in the scenario. At the end, this new anti-chaff jamming method is validated by Monte Carlo simulations.