Blood coagulation is a complex biochemical network that plays critical roles in haemostasis (a physiological process that stops bleeding on injury) and thrombosis (pathological vessel occlusion). Both up- and down-regulation of coagulation remain a major challenge for modern medicine, with the ultimate goal to correct haemostasis without causing thrombosis and vice versa. Mathematical/computational modelling is potentially an important tool for understanding blood coagulation disorders and their treatment. It can save a huge amount of time and resources, and provide a valuable alternative or supplement when clinical studies are limited, or not ethical, or technically impossible. This article reviews contemporary state of the art in the modelling of blood coagulation for practical purposes: to reveal the molecular basis of a disease, to understand mechanisms of drug action, to predict pharmacodynamics and drug-drug interactions, to suggest potential drug targets or to improve quality of diagnostics. Different model types and designs used for this are discussed. Functional mechanisms of procoagulant bypassing agents and investigations of coagulation inhibitors were the two particularly popular applications of computational modelling that gave non-trivial results. Yet, like any other tool, modelling has its limitations, mainly determined by insufficient knowledge of the system, uncertainty and unreliability of complex models. We show how to some extent this can be overcome and discuss what can be expected from the mathematical modelling of coagulation in not-so-far future.