The Rayleigh wave exploration method is a commonly used near-surface exploration technique，and dispersion curve inversion is one of the key steps in the data processing of Rayleigh wave exploration. As a typical multi-parameter，multi-extreme，and highly nonlinear geophysical optimization problem，accurate and efficient inversion of dispersion curves is of great significance for calculating near-surface S-wave velocity fields and obtaining stratigraphic structure information. We propose a method for the Rayleigh wave dispersion curve inversion based on improved dung beetle optimizer（DBO）algorithm. The method uses Halton sequence to initialize the position of the population，so as to better control the spatial distribution of the initialized population，and it adopts different search strategies for different sub-populations by population division，so as to avoid the inversion search falling into the local optimum and realize the fast convergence of the algorithm. In this paper，three theoretical geological models and a set of practical data are used to verify the effectiveness of the improved DBO algorithm applied to the inversion of dispersion curves to obtain the subsurface S-wave velocity distribution. The results show that the new method is more effective and stable than the current mainstream improved adaptive genetic algorithm for dispersion curve inversion and can converge to the optimal solution quickly.