The water oil relative permeability curve has a great effect on the rules of water cut increase and production variation. It is one of the most important data in reservoir development. With regard to a reservoir with a high degree of heterogeneity, the flow properties are various in different positions of the reservoir. Therefore, neither a single average relative permeability curve for the whole reservoir nor different curves for different sedimentary facies can precisely describe the reservoir flow characteristics, which will cause great difficulties for the remaining oil prediction and potential tapping. Therefore, it is of great importance to build a prediction model for the water oil relative permeability curve, which can provide a calculation theory of the relative permeability curve for reservoir simulation using different relative permeability curves in different grid cells. The existing prediction models for the relative permeability curve have established the correlations between petrophysical properties and end-point values of the relative permeability curve and between end-point values of the relative permeability curve and the relative permeability curve independently. However, the relationship between petrophysical properties and the relative permeability curve has not been developed. For this reason, it is impossible to achieve the spatial distribution of the relative permeability curve according to petrophysical properties. Furthermore, the end-point values of the relative permeability curve are usually calculated directly by petrophysical properties, which may result in all predicted relative permeability curves shifting to the left or right unrealistically compared to the reservoir average relative permeability curve. As a result of the above-mentioned problems, taking into consideration the correction effect of the average relative permeability curve on predicting relative permeability curves in different positions of the reservoir, this paper obtains a correlation between petrophysical properties and the relative permeability curve on the basis of the statistical analysis technique and normalization method of the reservoir average relative permeability curve. Finally, a prediction model for the water oil relative permeability curve is established. A test based on the basic data of the relative permeability curve is performed to verify the effect of this model. The test result shows that the prediction procedure of this model is quick, easy, and reliable. It also indicates that the spatial distribution of the relative permeability curve satisfying the migration rule can be generated from petrophysical properties, which provides a basic calculation theory of the water oil relative permeability curve for reservoir simulation using different relative permeability curves in different grid cells.