Abstarct: Structural optimizations which include conditions and available limits or constrain function have been always one of the engineers and designers object. So far, many researches have been constructed on optimization and a lot of different ways of solving it have been suggested. Totally, methods of optimization can be divided into 3 basic types, mathematical method, optimality criterion and mexta-heuristic. In this research we use Genetic Algorithm which is one of the mexta-heuristic-method, in regard to shape optimization of 3D elastic structures. If we want to use this method, having objective functions and optimal constrains into design variable is needed. In this thesis we use finite element (FEM) for construction analysis. In this method construction is divided into some hexagonal brick elements. Therefore, for changing geometry in shape optimization, we should verify the coordinate of element nods and analysis the problem again. In this research weight or surface are considered as objective function and stress limits and displacement of shapes as constrains of optimization problem, respectively. On the other hand optimization in two directions has occurred at the same time. We bring some 3D example with validation examples in this thesis, which is provided from computer program.