Optimization of structure`s topology that provides minimal material consumption, weight and good heat-moisture performance is one of the most challenging task in today’s engineering science. Plywood structures provide a rational usage of wood (Sliseris and Rocens 2010) and one of the reasons is less sawdust in manufacturing. The typical hot-pressed plywood sheet’s maximal thickness is limited and in many cases there is insufficient bending stiffness. The object of research is plywood composite material that consists of top and bottom plywood flange and middle core that is formed from plywood ribs and foam-type material. As a result of structure complexity (many variable parameters) and highly material nonlinearity it’s topology optimization is very challenging problem. Therefore classical optimization methods, like response surface method may fail or find local optimum (Rao 2009). The optimization of plywood macro-structure`s topology is done by non-traditional optimization methods- Neural network (Yao 1999), Genetic algorithm (Goldberg 1989, Sliseris and Rocens 2011) and Simulated annealing method (Laarhoven and Aarts 1987). All of these methods are developed by inspiration of nature. The numerical simulation of plywood structure is done by classical (plane-stress) composite material method and Finite element method. The multi-objective structure optimization is done and minimized objective function`s are ratio of structures bending load bearing capacity and structure`s weight, heat conductivity and moisture caused structure`s expansion-shrink coefficient. The typical constrains are used- maximal stress criteria, maximal deformation criteria. As a result there are proposed an effective multi-objective optimization methodology for plywood composite material. For the typical cases there are obtained the rational topology of macro-structure.