Agricultural robot plays an important role in accelerating transformation of agricultural modernization and achieving intelligent agriculture. Field positioning and navigation technology is the foundation for ensuring the safe and efficient completion of various agricultural tasks by robots, and combining high-precision positioning information of robots to achieve efficient planning of work paths is the technical core of field positioning and navigation. A path planning algorithm based on macro-micro integration was proposed. Firstly, the algorithm generated a global static work path based on the macro mapping information of the operation area. While implementing robot operations, various radar sensors were used to dynamically monitor the micro work environment and path information of the robot in real time. Finally, path tracking algorithms such as MPC were applied to real-time process local and global work environment information to achieve real-time job path optimization and correction to ensure the smooth progress of field work. Experimental verification showed that when the adjustable distances on both sides of the robot during field operations were 0.2m, 0.1m respectively, the algorithm can reduce the average crop compaction rate during the operation process from 3.4058% and 1.2763% to 0.6772% and 0.1889%. Meanwhile, the algorithm improved the precision of field operations under the requirements of precision agriculture and had important significance for achieving the goal of high yield, efficiency, and quality in agriculture.