Aiming at the position estimation error caused by far-point extrapolation in the existing ballistic extrapolation algorithm of the gun reconnaissance radar, a ballistic extrapolation algorithm based on the inverse extended Kalman filter is proposed.Based on the traditional ballistic extrapolation algorithm, a filtering model of seven-dimensional state vectors including the trajectory coefficients is established in the algorithm, and the end point of forward filtering is regarded as the start point of inverse filtering to implement the inverse extended Kalman filtering.Starting from the end point of the inverse filtering, the fourth-order Runge-Kutta equation is used to extrapolate towards the ballistic starting point.Simulation experiments and results show that the extrapolation accuracy of the proposed ballistic extrapolation algorithm based on the inverse extended Kalman filter is improved by about 50% compared with that of the original algorithm.