Despite several decades of study, there is no rigorous method for scaling flow through heterogeneous permeable media. Existing methods neglect heterogeneity and spatial correlation, which we now know to be of fundamental importance in many processes. On the basis of statistical techniques and inspectional analysis, we present a general method to scale flow through heterogeneous permeable media for an immiscible displacement of oil by water in a two-dimensional, anisotropic, heterogeneous cross section with statistically stationary properties.Our results illustrate the interplay among local heterogeneity (the variance at a lower cutoff in a power-law variogram), global heterogeneity (the variance at an upper cutoff) and zonal heterogeneity (the power-law upper cutoff and exponent). The effects of the scaling groups obtained from the method have been examined in detail, and some important flow characteristics, which were unknown or poorly understood before, have been revealed.