Abstarct: Today, the concept of optimization, as a basic principle in engineering design issues, is accepted. One of categories of optimization in structures is size optimization of structural members to obtain the least weight of structure subjected to design constraint. Genetic Algorithm (GA) is one of the mexta heuristic algorithms that has proved itself in this field. But the slow rate of optimization process through GA and also accidental start of optimization process by GA in some cases leads to a local optimum and stop there. In this thesis, it is tried to improve GA by combining one of the mathematical optimization methods (the method of Sequential Linear Programming) and Genetic Algorithm. To achieve this, by using the existing rules in the algebric method of forces in analyzing structures, a new formulation in the field of optimization of cross section of truss structures is presented. The point to be mentioned is that before starting the optimization process by GA, by using the method of Sequential Linear Programming, the area of optimum answer is determined. Then, Genetic Algorithm is used to find the optimum. In this method, mathematical optimization method is utilized to make the primary population of GA and hence the quality of the results is improved.