n order to real-time achieve accurate sliding parameters between tracks of a agricultural tracked robot and ground and improve control accuracy, a dynamics equation and nonlinear measurement equation of agricultural tracked robot were deduced by analyzing working principle of an agricultural tracked robot and using dynamics principle. Then an estimation system based on the unscented Kalman filter(UKF) was designed to estimate the precision pose parameters of agricultural tracked robot. Through this estimation the system sliding parameters might be reconstructed relying on kinematics analysis, which couldn`t be directly measured before. The simulation and experimental results suggested that with careful modeling of agricultural tracked robot, the estimation system was able to provide reliable and high update rate sliding parameters, which were taken as the basis for accurate control.