Objective To develop simulation methods that can evaluate insulin therapy protocols for glucose management in hospitalized patients with type 2 diabetes, in advance of clinical trials Methods We proposed simulation method using a virtual patient model that can evaluate insulin therapy protocols for glucose control of hospitalized patients with type 2 diabetes. For validating the method, we employed and compared insulin protocols of the clinical trials in a computational environment: sliding scale regular insulin protocols (SSI) vs. basal bolus insulin protocol (BBI). We also compared the glucose profiles in simulation results with that in clinical results. Results Using the model, we created 100 virtual hospitalized patients with type 2 diabetes (91.6 ± 5.8 kg of body weight, 205 ± 9 mg/dL of initial glucose). The mean admission glucose was 212 ± 9 mg/dL. A glucose target of 100~140 mg/dL was achieved when treated with BBI1 compared with the SSI (32.9 ± 2.0 % vs. 14.5 ± 4.1 % of time, respectively; P < 0.05), without increasing the risk of hypoglycemia. Each glucose profiles were comparable to those of the previous clinical study. Conclusion We demonstrated validation of the simulation methods. We showed that the indices for assessing efficacy and safety of the BBI were superior to the SSI. We also proved that the overall changes in daily mean glucose levels were in close agreement with clinical study. The simulation method could be a useful to evaluate insulin protocols before clinical trials without risk of danger and cost.