Abstract: With the wide use of unmanned planters, there are more and more collision events between unmanned planters and workspace. It is of great significance to reduce the security risk of seeding scene to avoid and reduce the space collision of unmanned planters. It provides an automatic path planning algorithm for unmanned planter based on RRT deep reinforcement learning algorithm. The optimal path planning fully considers the obstacle avoidance strategy, which can effectively reduce the probability of space collision accidents. And it has practical application value for the safe operation of unmanned planter.