Disturbance rejection rate (DRR) not only affected output accuracy of the seeker but also worsened the performance of missile guidance system. The main factors which caused the problem of DRR with semi-strapdown imaging seeker were scale error and disturbance torque. Firstly, a model of guidance system which contained disturbance rejection rate parasitic loop (DRRPL) was established. The relationship between DRR caused by different factors and the stabilization of guidance system was analyzed on basis of the model. Then, an online rejection method based on strong tracking unscented kalman filter (STUKF) was proposed to solve the problem above. The nonlinear filter model considering DRRPL was built by working principle of semi-strapdown seeker, the STUKF algorithm was adopted to estimate DRR and real line of sight rate. Finally, a math simulation was put forward. The result indicates that the guidance accuracy is improved after the compensation of DRR, and the compensation method had good anti-interference and robustness. The above theoretical analysis can provide guidance for online identification and compensation of semi-strapdown seeker disturbance rejection rate.