<p>This work aims at the integrated scheduling for whole refinery processes with consideration of uncertainties in demands and yields. We first define the scheduling problem for refinery processes. The external and process uncertainties of refineries are represented through the extended scenario tree based on historical data. Using a stochastic programming model, a hybrid Mixed-Integer Nonlinear Programming (MINLP) and Generalized Disjunctive Programming model is formulated for the scheduling of whole refinery processes under scenarios involving uncertain demands and yields. To solve the proposed stochastic programming model, we develop an Outer-Approximation method combined with the Fix-and-Relax strategy to efficiently solve real scheduling instances. The computational results demonstrate the value of the stochastic solution of the proposed scheduling model and the efficiency of the proposed solution method for larger number of scenarios compared with the MINLP solver DICOPT.</p>