We present an inverse scattering problem for recovering the shapes of multiple conducting cylinders with the immersed targets in a half-space by the genetic algorithm. Two separate perfect-conducting cylinders of unknown shapes are buried in one half-space and illuminated by a transverse magnetic plane wave from the other half-space. In the shape expansions, the cubic-spline method is utilized to describe the shapes of objects. Based on the boundary condition and the measured scattered field, we have derived a set of nonlinear integral equations, and the inverse scattering problem is reformulated into an optimization problem. The improved steady-state genetic algorithm is used to solve the global extreme solution. Here, frequency dependence on the inverse problem of buried multiple conductors is investigated. Numerical results show that the reconstruction is good in the resonant frequency range, even when the initial guesses are far different from the original shapes. On the contrary, if the frequency is too high or too low, the reconstruction becomes bad. In addition, the reconstructed errors for different distances between two conductors are investigated. It is found the reconstructed results are poor when the distance between two conductors is less than about a wavelength.