Dynamic nonlinear models of metabolism offer significant advantages as compared to traditional steady-state counterparts for the analysis of metabolic networks and identification of metabolic engineering strategies. Uncertainties in the metabolic network structure, kinetic rate laws, and their corresponding parameters are limiting the development of systematic methodologies for construction of large-scale, dynamic metabolic models. In this study, we proposed a novel methodology for design of metabolic engineering strategies that makes use of these models. We illustrated this methodology using a dynamic model of E. coli composed of 225 reactions and 140 metabolites. Implementation of this methodology and subsequent analysis allowed us to identify globally and locally relevant design strategies to from the current state of metabolic network transition towards a desired state. Using dynamic simulations, we analyzed the network responses upon the chosen perturbations of the network components to explore the feasibility and properties of the identified strategies. The methodology can be used for the design of any metabolic network of interest irrespectively of the organism and/or the physiological conditions.