For the obstacle avoidance control of agricultural tractor-trailers, a controller based on nonlinear model predictive control was proposed. Since the obstacle avoidance control of the agricultural tractor-trailer should take the position state and the attitude state of the tractor and trailer into account, the kinematics model of the agricultural tractor-trailer was derived based on rigid body kinematics and non-holonomic constraints. And based on this kinematics model, the position state and the attitude state prediction model of the agricultural tractor-trailer was established. Thereafter, the optimization function was designed based on the improved forbidden zone penalty function, and the proposed obstacle avoidance controller design was completed. In the simulation results, by comparing with the obstacle avoidance controller that did not consider the position state and the attitude state of the trailer, it can be known that the proposed obstacle avoidance controller of agricultural tractortrailers can control the tractor and the trailer to avoid the obstacle simultaneously. And under the control of the proposed obstacle avoidance controller of agricultural tractor-trailers, the minimum distance between the trajectory of the end point of each axle and the center of the obstacle minus the sum of the obstacle radius and the safety margin was not less than 0m.