To solve the problem of global navigation satellite systems and inertial measurement unit fusion time asynchronous and improve the accuracy of pose estimation of plant protection UAV, this article proposes a delay pose compensation algorithm based on the improved error state Kalman filter by using the characteristics of large inertia and strong vibration of plant protection UAV. Firstly, the nominal state variables are linearly predicted, and a fading factor is introduced to improve the system stability in strong vibration environments. Then, complementary filtering is used to compensate for diagonal velocity and correct the attitude error state variables. Finally, combined with the delay time measured, complementary filtering is used to extrapolate the data and improve the velocity and position accuracy under high inertia characteristics. Experimental results show that, compared with the error state Kalman filter algorithm, the root mean square error of roll angle and pitch angle is reduced by 0. 266 9° and 0. 241 4°, and the root mean square error of yaw angle is reduced by 0. 076 4°. Under normal track plant protection operation, the root mean square error of velocity in the northeast sky direction values are decreased by 0. 210 5, 0. 184 9, and 0. 238 8 m/ s. The root mean square errors of the northeast celestial position are reduced by 0. 21, 0. 19, and 0. 23 m, respectively. The algorithm effectively improves the accuracy of pose estimation.