Abstract: Due to high computational complexity of the traditional RRT path planning algorithm and the limited computational resources of the embedded platform, the real-time performance of the system is not effectively guaranteed. In order to improve the real-time planning of the unmanned system based on the embedded platform, a RRT path planning method based on state grid was proposed. Firstly, the state grid algorithm was used to generate the search extension domain to guide the RRT algorithm, which guided the search direction and reduces the path computational complexity of the RRT algorithm. Next, the random sampling point generation function was introduced to limit the area of sampling nodes. Then, the excessive steering paths were pruned to further achieve smooth paths by the B-sample curve function, which could improve the real-time performance of unmanned vehicle path planning. Finally, the embedded platform-based unmanned system was built for experimental test. The experimental results show that the total path length of the random search tree is reduced by 14%, the average path length after pruning is reduced by 16%, and the average time is reduced by 64% compared with the traditional RRT algorithm. Compared to the fusion algorithm of RRT and artificial potential field, the total path length of the random search tree is reduced by 8%, the average path length after pruning is reduced by 6%, and the average time is reduced by 33%. Therefore, the proposed algorithm has a smaller number of iterations, and can reach the target point faster while ensuring the continuity of the path curvature.