Aiming at the problem that standard RRT algorithm[Rapidly exploring random tree] using pseudo-random sequence leads to uneven and unreasonable distribution of sampling points, and there are redundant sections and redundant nodes in the path of mobile robot from the starting point to the target point, HDRRT[Halton && Dijkstra && Rapidly Exploring Random Tree] algorithm is proposed, which uses Halton sequence with good uniform distribution of sampling points for sampling, and uses candidate point set strategy to filter nodes to eliminate redundant nodes. At the same time, the improved Dijkstra algorithm is used to extract the key nodes of the original path to reduce the redundant path sections. On this basis, the path is smoothed by cubic B-spline curve. The simulation results of ROS system combined with MATlab show that HDRRT algorithm has the advantages of rapidity, stable planning and smooth path compared with Bias-RRT and standard RRT algorithm.