A methodology to test the possible deviations of fracture relative permeabilities from straight-line relationship with saturation is discussed. It is shown that with matrix relative permeability data known, it is possible to obtain fracture relative permeabilities by history matching experimental data with double porosity numerical models output. Rather than linear saturation dependence, fracture relative permeability is a power law function of the corresponding phase saturation. The impact of several important parameters like double porosity formulation, capillary pressure and shape factor is discussed. It has been found that double porosity formulations (other than the double permeability model) could represent the flow in a fractured medium. (C) 2001 Published by Elsevier Science B.V.