Abstract: The grey wolf optimization algorithm (GWO) has some demerits on solving more complex nonlinear optimization problems, such as relapsing into local extremum, slow convergence velocity and low convergence precision in the late evolutionary. this paper proposed a kind GWO algorithm based on levy flight (LGWO) via introducing of levy flight that is random walk. When updating the wolf's position, levy flight was used to conduct a global search on the wolf to prevent the wolves from losing diversity and easily falling into a local optimal solution. Through simulation, compared with the classic cuckoo search algorithm、particle swarm optimization algorithm and other group intelligence algorithms, the results show that the grey wolf optimization algorithm based on levy flight can effectively improve the accuracy of the solution and convergence speed, and the optimization effect is better.