Modern logging tools, including Nuclear Magnetic Resonance (NMR), allow geoscientists to integrate data and use advanced petrophysics techniques. The NMR logs estimate permeability curves based on empirical equations and their response are sensitive to the permeability trends in rock layers of a heterogeneous reservoir. Therefore, it is possible to define flow units based on NMR flow capacity. This process can even be enhanced if there are Production Logging results available. This paper presents a technique applicable to well test interpretation which consists in estimating the permeability of flow units through the integration of NMR, formation tests and production log data. The properties of each flow unit such as thickness, average permeability and effective porosity are estimated from the NMR log. Using these data a multilayer flow model is built, in order to determine a multiplicative factor for the prior defined average permeability from each flow unit honoring the results of DST data. Several simulations are performed by changing the horizontal and vertical permeabilities and skin factor for each flow unit, to obtain compatible results with the well test. This technique is applied in Brazilian carbonates. The cumulative NMR permeability curves shows good agreement with PLT data, which means that the NMR can be used as a predictive flow unit identifier when PLT data are not available. In conclusion, the technique permits the interpreter to test various scenarios in a multilayer model, using different combinations of permeability and skin values for the flow units. This petrophysical model can also be used as an input for the dynamic simulation of the reservoir.