Abstract: To solve the problem that the gray wolf optimization (GWO) algorithm is easy to fall into the local optimum and the convergence precision is poor, this paper proposes an improved gray wolf optimization algorithm based on opposition search and Levy flight strategy(OLGWO). In the initial stage of the algorithm, the opposite search strategy is adopted to narrow the range of feasible solutions. And in the process of updating location of the gray wolves, Levy flight strategy is adopted to avoid the algorithm falling into the local optimum. By four standard test functions, simulation experiments show that the proposed OLGWO algorithm is superior to the GWO algorithm in terms of convergence speed and solution accuracy, and it can search for the optimal solution quickly and accurately. Next, based on the OLGWO optimization algorithm, the ray tracing propagation model of the tunnel is corrected. The results show that the corrected model has better performances in terms of RMS and linear correlation and can achieve precise prediction of signal receiving power in the railway tunnel environment.