Models of geological objects are tools for interpolation and extrapolation of available data in space and time continuously. Real structures of the objects are unknown, and their models and simulated results carry uncertainty which cannot be evaluated in a provable way. The real issue is obtaining effective predictions in a reasonably defined sense. This requires a knowledge of mechanisms that convert actual geological properties into effective model parameters. These mechanisms are introduced in the book. They reveal that effective parameters are not statistics but characteristics optimizing the system made up by geological surroundings, their models, predictive problem formulations, including mathematical models of the simulated processes, boundary conditions, monitoring networks, criteria of efficiency and even by time. Examples of evaluating and applying transformation for assigning effective parameters and solving inverse problems are presented.