This paper mainly studies the problem of how to realize the optimal parking decision, and puts forward the corresponding calculation and analysis methods for the parameters of turning radius, turning lane width and minimum acceleration distance of unmanned vehicles. Firstly, according to the principle of Ackermann's steering geometry, the reasonable Angle of front inner wheel and front outer wheel are designed, and the minimum turning radius and minimum turning lane width are calculated. Secondly, based on the knowledge of kinematics, the shortest distance required for the vehicle to accelerate to the maximum limit speed when driving in a straight line is calculated. Then, the vehicle kinematics model is established, and the path planning methods of vertical parking, parallel parking and oblique parking are proposed. Finally, from the perspective of unmanned vehicle, the main factors affecting parking selection are determined, and the optimal parking decision-making model is established by using grey entropy theory, and the optimal parking space is finally determined. In addition, the results show that the algorithm has a small complexity and is suitable for any occupied parking space. The research of this paper provides useful reference and guidance for the path planning and parking decision of unmanned vehicles.