A fundamental development direction in structural analysis is driven by increased incorporation of computational intelligence. This contribution provides an overview of incorporation of computational intelligence in the broad research fields of civil and mechanical engineering, especially focusing on the objective of realistic safety assessment by efficient simulation with comprehensive uncertainty modeling. The effort of experimental testing in development of composite materials can be massively reduced by multi-scale simulation methods. The replacement of experimental testing by numerical labs is based on the complex numerical simulation of materials on the mesoscale in order to incorporate the material characteristics on the macroscale, i.e. for the simulation of complex load-bearing structures. This such frameworks of homogenization, uncertainty of the data and models are required to be taken into account for realistic safety assessment. The relevance of computational intelligence for solving such current problems in research is presented using the example of numerical homogenization for concrete material – consisting of cement matrix and aggregates. It is demonstrated that realistic engineering problems, involving realistically modeled uncertainty, require innovative homogenization approaches that are only possible by support of (artificial or) computational intelligence.