Physiologically realistic results are the aim of every blood flow simulation. This is not different in aorto‐coronary bypasses where the properties of the coronary circulation may significantly affect the relevance of the performed simulations. By considering three patient‐specific bypass geometries, the present article focuses on two aspects of the coronary blood flow – its phasic flow pattern and its behaviour affected by blood rheology. For the phasic flow property, a multiscale modelling approach is chosen as a means to assess the ability of five different types of coronary boundary conditions (mean arterial pressure, Windkessel model and three lumped parameter models) to attain realistic coronary haemodynamics. From the analysed variants of boundary conditions, the best option in terms of physiological characteristics and its potential for use in patient‐based simulations, is utilised to account for the effect of shear‐dependent viscosity on the resulting haemodynamics and wall shear stress stimulation. Aside from the Newtonian model, the blood rheology is approximated by two non‐Newtonian models in order to determine whether the choice of a viscosity model is important in simulations involving coronary circulation. A comprehensive analysis of obtained results demonstrated notable superiority of all lumped parameter models, especially in comparison to the constant outlet pressure, which regardless of bypass type gave overestimated and physiologically misleading results. In terms of rheology, it was noted that blood in undamaged coronary arteries behaves as a Newtonian fluid, whereas in vessels with atypical lumen geometry, such as that of anastomosis or stenosis, its shear‐thinning behaviour should not be ignored.