Abstract: To mitigate the energy imbalance problem caused by multiple uncertain factors in a regional integrated energy system, this report proposes a bi-level optimal energy scheduling strategy that combines the information gap decision theory (IGDT) with the model predictive control (MPC). The upper layer employs the IGDT to model the uncertain factors in the day-ahead scheduling plan to reduce the large deviation between the day-ahead scheduling plan and the actual scheduling caused by the source and load prediction errors. The lower layer uses the MPC rolling optimization within a day to correct the day-ahead scheduling deviation and compensate for the shortcomings of the IGDT open-loop control. To support the IGDT in dealing with multiple uncertainties, this work implements a two-point estimation method and the Cornish-Fisher series expansion to transform the multiple uncertainties (in terms of the source and the load) into a single power deficiency uncertainty, which improves the solving efficiency of the IGDT model. Finally, the developed method is applied to a practical example to verify its effectiveness.