<p>Dispersion curve inversion is a key step in Rayleigh wave data processing, which can effectively obtain underground S-wave velocities. However, the inversion of dispersion curves has characteristics such as multi-parameter, multi-extremum, and nonlinearity, which leads to the complexity and uncertainty of the inversion process. Traditional nonlinear algorithms often suffer from complex algorithm structures and poor balance between global and local search on the basis of the classical Pigeon-inspired optimization (PIO) algorithm, the Osprey optimization algorithm (OOA), and the Cauchy mutation strategy intergraded to develop an improved Pigeon-inspired optimization (Osprey–Cauchy and Pigeon-inspired optimization, OCPIO) algorithm for Rayleigh wave dispersion curve inversion. OCPIO combines the balance mechanism and dynamic adjustment capability of OOA, as well as the ability of Cauchy mutation to escape from local optima, significantly enhancing the algorithm’s global and local search capabilities, and effectively avoid the probability of inversion falling into local optima. Additionally, we also utilized logistic chaotic mapping to enhance the randomness of the initial population. The effectiveness of our method was verified through the inversion of Rayleigh wave dispersion curves using two geological models and further applied to a real seismic dataset processing. The research results indicate that our algorithm has strong balanced search ability, effectively avoiding the situation of inversion falling into local optimal solutions, while significantly improving the accuracy and stability of inversion.</p>