The invention discloses a method for planning a shortest path of agricultural machinery based on a Dijkstra algorithm, and relates to a method for planning a shortest path. The method comprises the following specific operation steps of: constructing a grid map of an operation area of the agricultural machinery, the grid width being divided according to a vehicle body width; enabling the vehicle body width to be (2i+1) times the grid width according to an accuracy requirement and an actual need; dividing the grid map into a passable state and an impassable state according to unmanned aerial vehicle aerial survey and obstacle distribution fed back by a sensor; and in order to control the diffusion of optimized nodes in a reasonable range, introducing a maximum estimation value K. According to the method provided by the invention, a group of better nodes are selected and the maximum estimation K is introduced to limit the diffusion degree of the group of nodes, and the group of better nodes is taken as a vertex set of the Dijkstra algorithm, so that the search time is reduced, and meanwhile, a found path is guaranteed to be an optimal path rather than a better path.