Genetic Algorithm is employed for cable structures subjected to large displacements, stress and diameter constraints. Design solution is governed by nonlinear interaction of cable sag, amount of pretension and stiffness of cable with deformations under transverse loads. Geometrical nonlinear finite element method with total Lagrangian approach is integrated in the present work with genetic algorithm to evaluate optimum values of design parameters, i.e. diameter of cable and amount of prestress. Study is performed on steel and carbon fibre cable stays, cable nets and cable roofs. Different types of crossover operators are applied along with single- and multi-level optimization techniques in order to obtain true optimum in the search space of global optimum.