GNSS controlled (Global navigation satellite system) land leveling technology plays an important role in improving the landform of field surface, also in enhancing use efficiency of water and irrigation efficiency. Aiming at improving the working efficiency of GNSS controlled land leveling system, the application of real-time intelligent navigation technology, which provided decision-making aid in land precise leveling, had explored. A tension sensor was added into this GNSS controlled land leveling system to realize the load monitoring of the forklift. The model of real-time land planning and navigation during the time of land leveling was set up, using mathematical modeling methods. The destination point of tractor was calculated with the load data and the terrain around it, in the principle of reducing the times of overload and empty load situations. Therefore, the turning angle could be calculated to give the navigation information to the users. Farmlands experiments were taken in the end and it turned out that: firstly, this method could be used to guide the driving in the process of land leveling quite well. Secondly, within the same leveling time in these experiments, the leveling effects with land planning and navigation functions showed to be much better than those of without. Finally, the sum proportion of overload and empty load when the land planning and navigation functions worked was no more than 69%, much better than the situation of without these functions.