In operations research, bi-level programming is a mathematical modeling which has another optimization problem as a constraint. In the present research, regarding the current intense competition among large manufacturing companies for achieving a greater market share, a bi-level robust optimization model is developed as a leader-follower problem using Stackelberg game in the eld of Aggregate Production Planning (APP). The leader company with higher in uence intended to produce new products, which could replace the existing products. The follower companies, as rivals, were also seeking more sales, but they did not have the intention and ability to produce such new products. The price of the new products was determined by the presented elasticity relations between uncertain demand and price. After linearization, using the KKT conditions, the bi-level robust model was transformed into an ordinary uni-level model. Due to the NP-hard nature of the problem, Benders Decomposition Algorithm (BDA) was proposed for overcoming the computational complexities in large scale. Finally, using the real data of Sarvestan Sepahan Co as a leader company, the validity of the developed model as well as e ciency and convergence of the BDA was investigated. The computational results clearly showed the e ciency and e ectiveness of the proposed BDA.