The problem is applied to the tank farm of a real world case company, CLC – Companhia Logística de Combustíveis, which is located in Portugal. The company receives 7 products from a pipeline connected to a refinery and stores them in 33 tanks. Customers provide forecasts for their daily demand per product on a monthly basis, which is the time horizon used by CLC’s schedulers to plan pipeline and tank farm operations. This will be the time horizon considered in this work. The pipeline flowrate considered is 400 m3/h. At each day, 3 shifts are considered and the customers’ demands have different consumption rates that depend on the shift. Two scenarios of one month inventory at a real world case study (April and May 2011) were proposed to evaluate model performance. Since it was impossible to find a solution for the first two or more weeks of April, it was proposed to divide the month in periods with duration of less than two weeks and apply the Rolling Horizon strategy. The MILP model was implemented in GAMS 23.5, CPLEX 12.2, on an Intel Atom N280, 1GB RAM. The results for April and May 2011 are presented in Tables 1 and 2, respectively. The number of periods for the Rolling Horizon was evaluated, both on extremes ratio and CPU time observed for scheduling of the whole month.