The automotive suspension system can mitigate road impacts and reduce the jolts caused by road unevenness, playing a decisive role in the ride comfort and handling stability of the vehicle. Establishing an accurate equivalent model of the suspension system is the foundation for the development of suspension control strategies. In order to make the simplified suspension dynamics model better reflect the response of the actual structure, this paper proposes a method for parameter identification of the stiffness and damping characteristics of the suspension system based on the transfer matrix method of the multibody system and the improved black-winged kite algorithm. Firstly, a quarter-vehicle multibody dynamics model based on the MacPherson suspension is established in ADAMS, and the actual dynamic responses of the suspension are obtained through different random road surface inputs. Secondly, a dynamics model of the suspension system is established based on the transfer matrix method of the multibody system, and a program for calculating the dynamic responses is compiled. Finally, the improved black-winged kite algorithm based on chaotic mapping, opposition-based learning, and the fusion of multiple strategies is used to conduct parameter identification of the stiffness and damping characteristics of the suspension. Results show that, compared with other machine learning methods, the method proposed in this paper has significant advantages in terms of accuracy and efficiency in identifying the parameters of the suspension system. It can help establish a suspension dynamics model that is closer to the real model, which has practical significance for the development of the suspension system and the research on the vehicle’s overall control strategy.