Metabolism modeling has the potential to provide guidance on medicine manufacture and medical treatment in an efficient and economic way. A pancreas model developed from a mathematical description of the relevant physiology will demonstrate insulin dynamics and provide a platform for comparing metabolic abilities among different health conditions. As a part of whole body metabolism model, the pancreas model is composed of three compartments: insulin clearance (Id), generation (Ig) and release into plasma (Ir). Based on data from intravenous glucose tolerance tests (IVGTT) available in the literature, a deterministic optimal search method called DIRECT (Dividing RECTangles) was implemented to find the model parameters. Validation was performed on data sets that are different from those used for the optimization.