Abstract: In the current path planning for unmanned oilseed rape sowing operations with intelligent agricultural machinery, the planned operation paths often result in relatively low field turning efficiency of the machinery, and the path planning does not take into account the fuel consumption during turning and the mid-operation replenishment of seed and fertilizer boxes, which leads to poor overall operation economy. Therefore, in view of the problems faced by the above path planning, this paper proposes an optimal path considering mid-operation resupply (OPCMR). Firstly, considering that the machinery often adopts fish-tail and U-shaped turning models when changing rows in rectangular fields, and these models are further classified into different specifications based on the number of rows between operations, the fuel consumption curves of the machinery in different states during turning are measured first. Then, based on the maximum turning radius and working width of the machinery, the two turning models are constructed. Finally, a turning cost function is established with the turning time and fuel consumption of the machinery based on these two turning models as constraints. To address the problems of slow convergence and easy trapping in local optimal solutions of the traditional ant colony algorithm (ant colony optimization, ACO), the traditional ACO algorithm is integrated with the simulated annealing algorithm and the heuristic factor of the ACO algorithm is dynamically adjusted to obtain an improved ant colony algorithm (improved ant colony optimization, IACO). Then, based on the turning cost function and the IACO algorithm, the optimal operation path with the least turning time and fuel consumption can be generated. Considering the uncertainty of the field entrance position, the function of manually setting the field entrance is added in the path planning, and the first operation of the planned path will start from the position closest to the field entrance. Finally, based on the key parameters such as the capacity of the seed and fertilizer box of the combined seeding and fertilizing machine and the per mu sowing and fertilizing amount in the field operation, an autonomous seed and fertilizer replenishment control method for intelligent agricultural machinery is proposed. Based on this method, specific path interruption points can be generated on the optimal operation path, and connecting these points with the pre-set seed and fertilizer replenishment stations can generate specific seed and fertilizer replenishment paths. When the seed and fertilizer of the machinery are insufficient, it can go to the seed and fertilizer replenishment station along this path for replenishment and then return along this path to continue the operation. Simulation experiments show that the IACO algorithm reduces the number of iterations by up to 56.9% compared to the ACO algorithm and reduces the minimum turning cost value by up to 2.82%, effectively avoiding the problems of slow convergence and easy trapping in local optimal solutions of the ACO algorithm. The seed and fertilizer replenishment path tracking experiments show that the average distance between the machinery parking point and the seed and fertilizer replenishment station is no more than 0.5 m, and the average initial deviation when the machinery enters the next operation row is no more than 0.03 m, indicating that this seed and fertilizer replenishment control method can effectively complete the seed and fertilizer replenishment task without affecting the operation of the next row. The path effect comparison experiments show that compared with the traditional comb-shaped operation path, the unmanned agricultural machinery based on the OPCMR path has a 17.5% increase in turning efficiency, a 9.8% reduction in turning fuel consumption, and a 94.7% reduction in human resource input after operation. This indicates that the path planning method in this paper can effectively improve operation efficiency and reduce resource input. This research can provide technical support for the construction of unmanned oilseed rape farms.