Aiming at the lack of stability and slow convergence of rapidly random-exploring trees (RRT) algorithm, an improved RRT path planning algorithm is proposed. Firstly, based on bidirectional search path planning algorithm, an improved combination strategy is proposed. The combination strategy combines target point search strategy with random search strategy, which not only improve the efficiency of searching but also guarantees the fast speed and obstacle avoidance. Finally, the algorithm smooths the redundant points and creates a new smooth route. The simulation experiment results demonstrate the effectiveness and stability of the proposed algorithm.