The near space hypersonic gliding target (HGT) is characteristics of highspeed, highmobility motion. In order to improve its tracking accuracy, an improvement on the existing Jerk tracking model is made. In this paper, the idea of firstorder AR model is applied to the real-time estimation of model parameters. The relevant parameters are introduced in the target state equation and measurement equation. And then the parameter adaptive Kalman filter algorithm is constructed based on the improved Jerk model to make the target state, and the model parameters adaptively adjusted at the same time. Finally, based on the HGT simulation trajectory, the proposed model is compared with the traditional Jerk model, and the effectiveness and advantages of improving the Jerk model for HGT tracking are proved.