OBJECTIVES:  The aim of this study was to propose an new model to predict the high cost of patients with advanced liver disease in waiting list for liver transplant and to evaluate the multiple factors that lead the high cost of treatment. METHODS:  It was proposed a logistic regression model combined to cluster analysis to evaluate all costs and amount of resources used by 492 patients with advanced liver disease on waiting list for liver transplantation. A microcost analysis were performed to evaluate the total cost including, consumables, image exams and laboratorial tests, procedures, blood bags reposition, chemotherapy and lenght of stay in hospitalization. The patients were grouped by ethiology and resource utilization following overall cost and the main factors that causes the high costs were analysed. RESULTS: Of the 492 patients, 247 patients (50,2%) obtainded cost over US$ 8,000, 152 patients (30,9%) obtained cost between US$ 8,000 to US$ 20,000 and 89 patients (18,9%) obtained cost over US$ 20,000. Four patients (0,81%) obtained cost over US$ 72,000. Two groups were created with cluster analsysis methodology. Mann-Witney test shows statistical diference between groups related to total cost (p=0.002). Logistic regression was performed with most expensive group (50%) and less expensive group (50%) and considering cluster analysis group as predictive variable. The results obtained were OR 1.89 (CI-95% 1.21 - 2.94) (p=0.005). CONCLUSIONS:  Cluster analysis and logistic regression model can improve the management of patient with advanced liver disease, effectively predict the chance of high cost of treatment and can be an important tool in cost analysis and in health economics assessment. The main factors that increase the overall cost of treatment of patient with advanced liver disease was chemotherapy, HCV, MELD over 29, blood bags utilization and multiple hospitalization.