A new evolutionary approach using a Genetic and Swarming Intelligence based Hierarchical Culture Algorithm is developed to identify the structure and parameters of nonlinear block-oriented systems of Wiener and Hammerstein types. In this scheme genetic recombination, considered as a culture macro operation, is used to adapt the structures of the linear and nonlinear elements of the model, while Swarming Intelligence based learning considered as a micro operation, is resorted to estimate their parameters. The hierarchical feature results from model clustering resorted to cope with structural and behavioral properties. The ability of the proposed method to accurately model complex and highly nonlinear systems is illustrated by numerical simulation.