Aim. To develop predictive model for hepatic insufficiency in obstructive jaundice. Material and methods. Obstructive jaundice was modeled by the author’s method on 48 mini pigs, while morpho-functional features of erythrocytes were studied by using of INTEGRA Aura atomic force microscope (NT-MDT, Zelenograd, Russia). Histological specimens were stained with hematoxylin and eosin. Discriminant analysis was used to create predictive model for hepatic insufficiency. Results. Mathematical model of hepatic insufficiency prediction has been developed. Sensitivity and specificity of this model were 94.1% and 74.2% respectively. Total percentage of correct predictions was 81.3%. Conclusion. Severe obstructive jaundice contributes erythrocyte’s transformation from biconcave to dome-shaped followed by changes of its physical properties. Erythrocyte’s volume and activity of cytolysis enzymes are the most informative to predict hepatic insufficiency. Our model allows us to diagnose this complication at early stages and to correct pre-, intra- and postoperative therapy.