Abstract: Aiming at the problems of poor levelling effect and poor stability of hilly and mountainous agricultural machinery, an electromechanical omni-directional levelling agricultural work machine was designed, and an automatic levelling control method based on BP neural network PID was proposed. Firstly, the whole structure scheme of the electromechanical work machine was provided. The electromechanical work machine is mainly composed of an upper working platform, a middle electromechanical omnidirectional leveling system, and a lower tracked driving system. The electromechanical omnidirectional leveling system can achieve a maximum leveling angle of 20° in the lateral direction and 25° in the longitudinal direction. Then, the control principle of the electromechanical omnidirectional leveling system was presented. The electromechanical omnidirectional leveling system uses angle error leveling method for leveling control. A mathematical model of the servo electric cylinder system was established, which mainly includes the mathematical models of the servo motor and transmission system. Then, a BP neural network PID leveling control algorithm was designed, in which the weight coefficients of the BP neural network were dynamically adjusted based on the body pitch angle control error, and the optimal PID control parameters were output to improve the leveling speed, leveling accuracy, and leveling stability of the electromechanical omnidirectional leveling system. Furthermore, a BP neural network PID controller and a servo electric cylinder system simulation model were established in Matlab/simulink. An integrated mechanical model of the entire machine was established in Adams, and a Matlab/Simulink-Adams co-simulation platform was established. In order to compare the leveling control performance of the two algorithms, a classical PID control algorithm was used as a reference, and the simulation and experiment were conducted to analyze the performance of the two algorithms. The simulation results show that in the lateral leveling control with a tilt angle of 20° and the longitudinal leveling control with a tilt angle of 25°, the leveling control performance of the BP neural network PID control algorithm is better than that of the classical PID control algorithm. The average reduction in the lateral and longitudinal leveling rise time is 20.09%, and the average reduction in the leveling stabilization time is 18.99%. The steady-state error is controlled within 0.30°. Finally, static and dynamic tests were conducted on the physical prototype. The static test results show that the BP neural network PID control has a shorter levelling stabilization time, smaller maximum inclination overshoot, and lower steady-state error in both lateral 20° and longitudinal 25° levelling control than the PID control. Compared with the simulation results, the differences in leveling stabilization time were 0.6s, 1.0s, and the differences in maximum inclination overshoot were 0.3°, 0.5°. The differences in steady-state error were 0.14°, 0.12°, and the errors remained within a small range. The dynamic test results show that, the electromechanical working machine operates at a speed of 3 km/h on hilly terrain with uneven surfaces, when the angle of inclination of the fuselage changes significantly, compared to PID control, the BP neural network PID control reduces the average stabilization time for both longitudinal and lateral leveling by 0.5 seconds, resulting in a faster leveling speed. Additionally, the errors in longitudinal and lateral leveling angles are within ±1.5°, indicating higher leveling accuracy. The effectiveness of the electromechanical omnidirectional leveling system and the BP neural network PID leveling control algorithm was verified through experiments, which could effectively realize the automatic leveling control of the working platform and meet the levelling performance requirements for hilly and mountainous operations.. The automatic leveling technology for agricultural machinery can effectively solve the problems of poor work quality caused by uneven road surfaces, large slopes, etc. The electromechanical working machine designed in this paper explores the structure design and leveling control method of the omnidirectional leveling system for agricultural machinery, which has certain practical significance.