Abstract: With the comprehensive promotion of the rural revitalization strategy, the process of agricultural and rural modernization has been accelerating, and the transformation of agriculture to intelligence, precision and environmental protection has become the main trend of development. Among them, autonomous navigation of unmanned agricultural machinery has become a research hotspot in recent years. The autonomous navigation technology of agricultural machinery can effectively improve the accuracy and efficiency of agricultural machinery, and path tracking control is the core technology of autonomous navigation of agricultural machinery. Many scholars at home and abroad have carried out a lot of research, and satisfactory results have been achieved under relatively ideal working conditions. However, agricultural machinery also needs to operate in complex and changeable soil conditions and highly unstructured farmland environments, such as muddy paddy fields, rugged slopes, slippery grasslands, etc., in these scenarios, agricultural machinery will inevitably encounter various uncertain disturbances in the operation process, which will affect the accuracy of path tracking and even affect the stability of the control system. Therefore, the autonomous navigation control method of agricultural machinery should have good anti-disturbance ability and strong self-adaptability. In order to improve the path following control accuracy and anti-disturbance ability of four-wheel self-steering (4WSS) electrically driven unmanned agricultural machinery in complex farmland environment, a non-singular fast adaptive terminal sliding mode composite control method based on nonlinear disturbance observer was proposed. Firstly, considering the influence of drive wheel slip on the chassis motion law in farmland environment, a 4WSS chassis extended kinematic tracking error model of unmanned agricultural machinery including unknown sideslip speed and steering slip angle is established, and secondly, a nonlinear disturbance observer based on the chassis extended kinematic tracking error model is designed to estimate the external lumped disturbance with unknown slip parameters, and the disturbance compensation is introduced into the controller. On this basis, in order to slow down the jitter problem of nonsingular fast terminal sliding mode control, a nonlinear adaptive sliding mode approach law is constructed based on the exponential approximation law to ensure that the controller outputs the acceptable smooth control amount of the steering actuator, and the stability of the proposed composite controller is strictly proved. The results show that when the 4WSS unmanned agricultural machine tracks the U-shaped target path at 0.4 and 0.6 m \cdot s, compared with the traditional non-singular fast terminal sliding mode control method, the proposed method can reduce the standard deviation of the lateral error of the target operation path at different speeds to 8.1 and 9.2 cm, respectively. In spray mode, the corresponding standard deviation of lateral error is reduced to 5.3 and 7.8 cm, respectively. The proposed method effectively improves the robustness and accuracy of the path tracking control of 4WSS unmanned agricultural aircraft in complex farmland environment scenarios with good anti-disturbance capability and strong adaptive ability to meet the needs of agricultural production.