Aiming at the problems of high search nodes, long evaluation time, and too many large continuous turning angles in the traditional A* algorithm for global obstacle avoidance and path tracking, which lead to the instability of obstacle avoidance in actual robot work. A Facilitative Global Points A* Algorithm (FGP-Astar) is proposed, which is a global path planning algorithm based on an improved A* algorithm. Based on the bidirectional A* search strategy, the adaptive weight factor of the heuristic function is configured to prioritize approaching obstacles and assist in tracking, improve the node search efficiency, and avoid the intervention of invalid nodes. Bezier curve is used for path smoothing and key path points are extracted. At the local path planning level, this paper addresses the deficiencies of traditional dynamic window approach that easily fall into local optimal paths and cannot avoid complex obstacles. A new sub-function is incorporated into the evaluation function of the Dynamic Window Approach (DWA), which is based on the traversable area of obstacle distribution. Gazebo simulation is applied to ROS1 mobile robot platform, and a comparison experiment of path optimization for fixed point delivery is carried out. The simulation results show that FGP-Astar reduces the global path planning time by 55.63% compared with the traditional A* algorithm, and the number of nodes traversed in the path prediction is reduced by 65.79%. It demonstrates the rapid deployment of the algorithm in global path planning, which reduced computational load, and obtaining more reasonable paths. Subsequent research needs to increase sensitivity and responsiveness in dynamic obstacle avoidance.