OBJECTIVES: There are increasing publication on cost-effectiveness studies of insulin analogue but their underlying models are yet to be critically appraised. This study aimed to review modelling methods used in economic evaluation of insulin analogue. METHODS: The literature review were conducted according to PRISMA framework to identify models used in economic evaluation studies of insulin analogue in diabetes mellitus. Relevant studies published between 1^st January 2005 and 30^thSeptember 2015 were identified from major database (e.g. PubMed and Econlit) and HTA report (e.g. CADTH database). RESULTS: A total of 47 published articles were selected for final analysis. Overall, majority of models (n=35) were Markov based, mostly (n=33) IMS CORE diabetes model. Other modelling approach included simple mathematical model (n=6) and discrete event simulation (DES) model (n=6). The most commonly used DES model (n=5) was Cardiff Diabetes Model. CORE, Diabetes mellitus model (DMM) and Cardiff diabetes model were internally and externally validated. Other models only were internally validated except for one. Mathematical model estimated 1-year cost-effectiveness based on a single adverse effect in diabetes population. Markov and DES model have similar in long-term outcomes measurement. However, DES model predicted risk of the first occurrence but Markov based model simulated first and recurrent diabetes-related-complications. Outcomes from the models highly depended on the application of the assumptions in the models. CORE diabetes model able to capture most complications of diabetes especially compared to DES models. DMM and Grima model unable to capture hypoglycaemia risk. The majority of models were run over a lifetime period, used a payer perspective and had the capability for sensitivity analysis. CONCLUSIONS:  A set of health states and adverse events to capture impact of insulin analogue on diabetes patients must be accounted in a decision model. For example, macrovascular complications like myocardial infarction must be included when measured cost-effectiveness of insulin analogue.