The constantly rising food demand of a steadily increasing world population requires improvement in efficiency, competitiveness, and productivity of current meat and dairy production systems. Thus, robotics-based approaches have an important role to play, especially in dairy cattle farming, because the intensive grazing systems depend on numerous time-consuming and tedious operations required to be carried out to assure an optimal cattle feeding as well as utilization of forage resources. These operations range from data acquisition considering the amount and quality of available forage within the paddocks, to carrying out maintenance operations in order to safeguard high yieldwith required quality and availability of the forage throughout thewhole grazing season. This issue is addressed within the ICT-AGRI project i-LEED, in which one of the main tasks is to build accurate and feasible trajectories for a scouting and a maintenance robot to fully or partially cover the paddocks, as well as to reach only targeted spots previously located. This paper presents an original and fully operational method for trajectory planning by designing segments of clothoids while taking into account additional dynamic constraints, such as the steering rate capacity of the robot, its speed, and the maximally allowed transverse acceleration. In case of specific points to reach within a paddock, this approach is completed by morphological operations to first define regions and next rank them w.r.t. the minimal length of travel by solving the traveling salesman problem. Based on the kinematic and dynamic properties of the scouting and maintenance robots devoted to the i-LEED project, the performances of the proposed planning approaches are presented.