The model of the human glucose-insulin system plays an important role in several clinical treatment methods and protocols, like tight glycemic control of intensive care patients. The Intensive Control Insulin-Nutrition-Glucose (ICING) model is one of these protocols that was used for the development of the Stochastic Targeted glucose control (STAR) protocol applied as a standard of care in New Zealand and Hungary. The original ICING model uses an ordinary differential equations (ODE) for the description of the glucose-insulin metabolic system.Recent studies attempted the extension of the original ICING model with stochastic terms resulting a new stochastic differential equations model(SDE). By using the resulted of this new model (SDE) we may have the chance to improve the accuracy of ICING modeling (reduce the modelling error) which can results a better clinical treatment using STAR. In the study presented in this paper, the potential use and implementation of the stochastic version of the ICING model (SDE), was analysed and the modelling error was compared with the original version of the model using a large clinical data set including treatment records of 60 patients from Belgium, Hungary and New Zealand. The results show that the SDE model gives a smaller modeling error compared to the ICING model in most of patients.These results suggest that SDE model may be used to improve the prediction process of the blood glucose level of the ICU patients which will be an important step of the STAR protocol.