Agricultural flexible chassis (FC) is a kind of in-wheel motor driving electric vehicle applied for greenhouse. The FC is composed of 4 identical off-centered shaft mechanisms, which can be controlled independently. Through this kind of mechanism, the FC can achieve various motion types. However, it is difficult to maintain the linkage relationship between the 2 front off-centered shaft mechanisms when the FC is in front wheel steering motion, and hard to accomplish smoothly turning. In order to solve this problem, a coupling control strategy for linkage motion was proposed in this study. Firstly, the contour error of the 2 wheel steering angle was deduced based on the Ackermann steering geometry and the cross-coupling control principle. To reduce the contour error during steering, a fuzzy PID control algorithm was designed, which could realize parameters self-tuning. As the electromagnetic friction lock was controlled by pulse width modulation (PWM) signal, the duty cycle of the PWM signal had to be adjusted according to turning situation. Thus, a fuzzy logic method was then introduced to regulate the duty cycle. Namely, the PWM signal duty cycle can be in line with the steering wheel signal size and its change rate, and the motion of the electromagnetic friction lock can match the steering speed of the off-center shaft mechanism. In this way, the coupling control of the 2 front off-centered shaft mechanisms can be achieved. The control strategy was then verified through simulation based on MATLAB/Simulink. To further verify the effectiveness of the control strategy, the control program was loaded into the steering system hardware, and tests were carried out on hard surface road. According to the traditional vehicle steering test methods, step steering test, snake steering test and random steering test were conducted. The effects of front wheel steering under the steering angle allocation control method were compared with coupling control strategy. The simulation results demonstrated that the proposed control strategy was effective and feasible. The test results on hard surface road showed that the FC had a fast steering response in step steering test. The response time of step steering was 0.8s and overshoot was 1.3° under coupling control method during front wheel steering of the FC. The overshoot under allocation control was larger than coupling control and its fluctuation was notable. From the results of the snake steering and random steering, it was obviously that the steering angles of the left and the right front wheels had good tracking performance for their target angles, respectively. The opening and closing of electromagnetic friction lock can match the steering motion of the electric wheel well. The maximum and average following errors of 3 steering modes were less than those under the allocation control method. The maximum and average contour errors of the linkage motion were 1.2° and 0.6° for step steering, 1.1° and 0.6° for snake steering, as well as 1.0° and 0.5° for random steering, respectively. All these errors were also smaller than the allocation control. Under the coupling control, the maximum steering error between simulation and test was 2.2°, and the contour error trend of them was consistent. The simulation model was reasonable and effective. The contour errors of the allocation control in these 3 kinds of tests had more fluctuations and larger range of fluctuations than coupling control. The steering performance under coupling control strategy was obviously better than the allocation control method. The steering angles of the 2 front wheels had maintained Ackerman steering geometry well. The coupling linkage control strategy proposed in this paper has good effectiveness and feasibility. This research can provide references for steering control or other applications of the FC.