Abstract: Considering an impure reference signal in airborne passive radar systems, the interference, which is the matched result of scatterer echoes with multipath in the reference channel, degrades the target detection ability of space-time adaptive processing (STAP). In addition, the off-grid effects seriously affect the interference estimation performance. Aiming at the interference suppression problem in airborne passive radar with off-grids effects, an interference suppression method based on local search is proposed in this paper. The algorithm builds a global range-Doppler dictionary based on the interference sparse measurement model, from which a global grid point matching the interference is selected. Then, the global grid point is used as the center to finely divide the Doppler plane and construct a local Doppler dictionary, and the local grid point that is more suitable for the interference is selected. Finally, the range-Doppler image of the interference is estimated to achieve interference suppression and overcome the influence of impure reference signals on the STAP algorithm. The simulation results show that the proposed algorithm can reduce the loss of suppression performance caused by off-grid effects and improve the target detection ability of space-time adaptive processing with impure reference signals.