Because of the existing problems in path planning of greenhouse robot under complex environment, a directional A* algorithm was proposed. This method was focused on the smooth design, collision detection and the algorithm efficiency. Firstly, the “line of sight” solutions were used to smooth the path for getting rid of the zigzag effect and collisions. Secondly, the “arc-line-arc” turning methods were applied to avoid the width of the greenhouse robot in path finding. At last, some basic optimizations based on the binary heap were carried out to speed up the directional A* algorithm. Simulation and comparison results between the improved A* algorithm and traditional one showed that the proposed method was more efficient. It can not only meet the requirements of smooth, but also predict collision after proceeding with turning strategy. At the same time, the accelerating algorithm based on the binary heap made the path finding 4~7 times faster. Moreover, a path planning and tracking test was carried out in laboratory environment, where a simulation greenhouse was built. The results verified that the tracking precision can keep in a small range and the greenhouse robot can run without collision when the navigation path was given by the proposed algorithm, which proved the effectiveness and feasibility of the directional A* algorithm.