This paper address the problem of motion control of an agricultural robot during the whole process of farming work; including maneuvers to follow straight crop rows inside the field, and operations of lane change at the headland. The navigation sensors consisted of an inertial measurement unit and a real-time kinematic GPS. The control algorithm was able to identify the bias of the yaw angle and compensate lateral deviation caused by inclination of the robotic tractor, in order to guide the robot accurately in the field. A circle-back turning algorithm based on continuous primitives connected together was used at the headland to create reference paths. In addition, steering maneuvers for lane change were performed while the robot tractor is moving in order to protect the surface of the soil at the headland. To follow the path accurately, the slide movement of the tractor and the steering rate were taken into account by estimating the turning radius on real-time. Therefore, guidance paths could be optimized based on the specific condition of the field. Finally, the performance of proposed autonomous maneuvers was evaluated in field experiments.