In this study, we demonstrate the use of combined inverse metabolic engineering and metabolic evolution for rational strain development. We have applied these two approaches for designing the efficient E. coli strains for the production of primary metabolites such as ethanol as well as for the production of complex secondary metabolites like lycopene. Inverse metabolic engineering based on Elementary Mode Analysis was employed to predict a set of multiple-gene deletion required for the cell to function according to an efficient pathway. Metabolic evolution of growth-based selection was used for co-selection of an increased production of the product. We have recently constructed the design strains and characterized their performances. Experimental results have shown that the designed strains achieved a better performance than the wild type under identical conditions. The results, therefore, confirm the validity of these applied strategies for rational strain development.