Iterated Function Systems are popular techniques for gene- rating selfsimilar fractals. An important practical problem in this field is that of obtaining the IFS code which approximates a given image with a certain prescribed accuracy (inverse IFS problem). In this paper we present an hybrid evolutive-genetic algorithm to solve the inverse IFS problem in two steps: First, an Evolutive Strategy (ES) is applied to identify a set of affine transformations associated with selfsimilar structures within the image. Then, the best adapted transformations are combined forming an initial population of IFS models and a Genetic Algorithm (GA) is used to find the optimal IFS model. We show that this hybrid algorithm performs significantly better than one-step global evolutive or genetic algorithms which have been recently reported in the literature.