In the case of development of a petroleum reservoir there is essential to have a suitable reservoir model to proper forecast of its future and studding on the enhanced oil recovery plans. The more the heterogeneity the harder to model the reservoir. Relative permeability is one of the most important parameters, influencing on fluid distribution and its mobility in the porous media. Relative permeability curves are determined from coreflood experiments. One of the fundamental problems in petroleum reservoir simulation is to estimate these curves all over the reservoir. In this work, the available methods are studied, and a new method of estimating relative permeability curves to reservoir scale that is based on geostatistics is introduced and examined; also two scaleup methods including capillary equilibrium and viscose limit was examined. After that, static models ranking methods as well as future reservoir behavior were examined with the 3000 Monte Carlo simulated models. The results shown that the difference between cumulative oil production of the new method and other ones depends on the variance of the input data. Although the water break through time of this method is comparibale with End Point Scaling Method and Reservoir Rock Typing Method. Also the effect of the number of input data on the precision and relative behavior of methods has been studied. In the case of replacement of the injection and production wells the new method does not show significant reaction.