Abstarct: In the last decades, the necessity of time and cost reduction in designing structures and economical consumption of their constituent materials has been felt increasingly. These necessities have given special importance to structural optimization problems. In the meantime, the optimal, rapid, safe, and economical design of truss structures -one of the most commonly used structures in construction industry- has received special attention. On the other hand, in most researches done, random nature of structural parameters such as load and resistance and their correlations have not been considered, and optimization has been done in a deterministic way. The goal of this research is to introduce a frxamework for simultaneous geometry and sizing optimization of truss structures, so that the total weight of structure obtains its minimum possible value under the structural reliability system constraint. In addition, this frxamework will increase the safety rating of structure to its target level. Using this frxamework, truss structures can be designed without applying simplifying assumptions and considering design code factors and requirements. Random nature of loads acting on structure and the resistance of its constituent materials are included in design process of structure by means of structural reliability theory. Also, use of structural reliability system constraint causes interactive effects of truss members to be considered in its designing. Structural system reliability analysis is done with branch and bound method – one of the failure-path-baxsed methods. However, most researchers have estimated failure probability of structural system baxsed on failure probability of members. This approach leads to undesirable and uneconomical results in optimization of structure. Usually, system reliability analysis of a large-scale truss is very complex and time-consuming. Therefore, in the absence of a practical solution, the optimization of trusses using system reliability constraint will be impractical. In this dissertation, we suggest strategies to improve results of branch and bound method and to reduce the time of system reliability analysis. Optimization is done by genetic algorithm method.