Under the condition of narrow parking space, problems such as low success rate or slow planning speed are encountered in the existing parking path planning based on the hybrid A* algorithm. In order to solve the problems, an improved hybrid A* path planning algorithm was designed. By using the arc spline as the reference path and using the points on the reference path as the target point of the hybrid A* algorithm, the path to successfully park into the garage was searched. Batch simulation tests based on MATLAB were conducted according to different parking space widths. The results show that a significant improvement in the success rate of vehicles parking into the garage in a specific area can be achieved by the improved hybrid A * algorithm, and at the same time, a certain level of planning efficiency was attained. Finally, a joint simulation experiment was conducted based on Prescan, Carsim and Simulink, which verifies that the tracking requirement of the real vehicle experiment is met by the path planned by the designed algorithm.