In the oil supply chain, the refined petroleum products are transported by various transportation modes, such as rail, road, vessel and pipeline. The latter provides one of the safest and cheapest ways to connect production areas to local markets. This paper addresses the operational scheduling of a multi-product tree-like pipeline connecting several refineries to multiple distribution centers under demand uncertainty. A new deterministic mixed-integer linear programming (MILP) model is first presented, and then a two-stage stochastic model is proposed. The aim of this model is to meet depot requirements at the minimum total cost including pumping and stoppages costs. The efficiency and utility of the proposed model is shown by two numerical examples, which one of them uses the industrial and real data.