The need for integration between the various components of the supply chain and the SEOP (Sales and Operations Planning) are concepts widely known by the companies, however it is often difficult to align theoretical concepts to the real needs and processes of companies. The focus of this work is the operational planning of the logistics supply chain of petroleum and derivatives of an oil company. The company uses a deterministic linear programming model for development of the plan. The parameters´ deviations between real and planning data, which influence the plan, as international prices, volume of oil production, demand for some oil derivatives and availability of refinery units were monitored. After analyzing of these deviations, we used the linear programming model of the company to develop a range of sensitivities, feeding back the model, using the mean errors of the variables. Finally the information of real x plan x modified plan (planned sensitivity) with uncertainty are grouped. The results show that the modified plan considering the uncertainty of the variables through the historical average errors enables a more robust planning, where the result is no longer a deterministic optimal value and it presents itself as a good range of values.