The paper focuses on crude oil refineries and presents a methodology to determine dynamic models of prices of different oils and the derived distillates. It shows how to take into account stochastic fluctuations of quotations due to market volatility. These dynamic models can be used to forecast possible economic scenarios in supply chain problems. They allow removing the rather limiting hypothesis of fixed prices, which is often present in the literature. Indeed, the paper shows how even on short-term horizons the historical fluctuations of prices of crude oil and distillates can be quite significant. Eventually, it proposes a methodology to simulate the fan of distinct dynamic scenarios that apply to the solution of the supply chain problem. The proposed methodology is appropriate for both short-term (i.e. planning) and medium-term (i.e. scheduling) problems, which may cover time horizons from few weeks up to some months.