Abstract Performance predictions made by reservoir simulators are highly dependent on experimentally-determined relative permeability data assigned to simulation grid cells. There is often no apparent relationship between this type of data and a quantifiable property of cells, meaning that samples within a group of neighboring relative permeability curves do not have similar values of a common characterization property. On the other hand, it is easier to correlate primary drainage capillary pressure data to cell properties and create a petrophysical static rock type (PSRT) from a group of similar capillary pressure data. The common approach is to average the relative permeability data corresponding to each PSRT and input that into a simulator. In practice, relative permeability curves corresponding to a given PSRT would exhibit a significant scatter, especially in carbonate formations. This will cause the resulting average relative permeability data to be highly uncertain and poorly representative of true dynamic behavior of the cells. In this study, a new technique is proposed to reduce uncertainties associated with classical methods of assigning relative permeability data to grid cells of a simulation model. The idea is based on a robust and universal criterion for characterization of dynamic characteristics of rocks from laboratory-derived flow data to define a petrophysical dynamic rock type (PDRT). A PDRT is defined as a collection of rocks with similar True Effective Mobility (TEM) behavior. A TEM-function for analysis of relative permeability is defined as a product of relative mobility by the square of a mean pore radius, and is analogous to the well-known J-function for scaling capillary pressure curves. For a given PDRT, a special weighted average of corresponding relative permeability data, or average TEM data should be used as input to a simulator. In the latter case, the simulator should have the capability to back-calculate cells’ quasi-relative permeabilities using the inputted average TEM data. Moreover, we show that factors such as wettability and pore-scale heterogeneity can cause phase-dependent dynamic characteristics (i.e., different dynamics with respect to each fluid phase). Hence, different dynamic rock typing schemes should be applied to the flow of each fluid in system. A structured methodology is presented to identify PDRTs using TEM-function and prepare input parameters to simulators by analyzing relative permeabilities using the universal criterion developed in this work. Ultimately, special core analysis data from the Asmari Formation, a carbonate reservoir from one of the Iranian Fields, is used to verify and demonstrate the applicability of the presented technique.