Abstract：Rice is one of the major food crops in the world, weeds are one of the main reasons for the decrease of rice yield and quality. In traditional agriculture, the main weeding method is to use herbicides, which would not only cause environment pollution, but also cause weed resistance. At present, mechanical weeding, as a green weeding method, is a new technology, because of the possibility of seedlings injury, it is still in the development stage. In order to avoid seedlings, there is a lag between the adjustment of weeding components and the movement of fuselage. During the rectification period, the weeding components will inevitably cause damage to rice seedlings. Therefore, it is necessary and urgent to realize intelligent seedling avoidance control technology by perceiving the curvature degree of rice rows, differentiating the operation path of weeding components. To reduce the seeding injury rate of mechanical weeding, on the basis of obtaining the information of rice seedling belt, a control system for weeding alignment in rice field based on linear active disturbance rejection control is designed in this paper. Row alignment technique is a method to control the deviation of weeding component on the operation path. By using the structure of inner and outer sliding beams, the weeding component can avoid seedling and control weeding in the working path through the row alignment mechanism. A hydraulic control system is proposed and simulation is carried out by combining Amesim with Matlab. The simulation platform of hydraulic control system is constructed, and the controller is designed. The simulation comparison between the linear active disturbance rejection control (LADRC) algorithm and PID algorithm is carried out. The results of simulation show that the line alignment time of LADRC system is 0.1s less than that of PID system, and in the case of disturbance, the anti-interference of LADRC is better than PID control algorithm and has certain robustness. To verify the actual performance of the LADRC, field experiments are carried out. The results show that the traveling speed and working depth have a significant impact on the seeding injury rate, and the adjusting space has a mild impact. The optimal working parameters combination is that the traveling speed is 0.5 m/s, the adjustment space is 60 mm, the working depth is 20 mm,, and the seeding injury rate is 3.6%. As the performance comparison experiment shown, the average seeding injury rate with alignment control system is 3.9%, while the average seeding injury rate without alignment control system is 18.6%. The system can meet the requirements of mechanical weeding control, and can provide references for the precise control problems in paddy field environment.