This work focuses on the study of Petrobras, regarding the strategic planning of the Company s investments, from an integrated oil supply chain perspective. From one of the most widely used mathematical models in the Company, several strategic decisions of great importance are supported, so as to maximize its operating result over a time horizon of approximately 10 (ten) years. Based in current literature, developments are proposed and tested in the mathematical model. First, two-stage stochastic programming techniques are introduced, where investment decisions are represented by first-stage variables; and system s operation – from oil refining and sales to the entire logistics issue – by second-stage variables, after realization of the stochastic parameters. In a second step, decomposition techniques are applied to circumvent any large scale limitations. The results show that the stochastic model starts to reach these limitations in problems with 30 scenarios or more. On the other hand, despite the considerably greater computational time, the decomposed model was able to solve up to 80-scenarios problems, during the tests.