Objective Currently，there are many methods for constraint-based optimization modelling using whole-genome metabolic networks to predict the metabolic flux distribution.Most of these methods require rates of metabolite uptake and secretion，as well as a priori knowledge of biological parameters，including biomass and the highest level of ATP production.However，measuring the rates of metabolite uptake and secretion in complex multicellular organisms is very challenging and is further complicated by different tissues having different metabolic goals.This presents a significant obstacle in modelling the metabolism of higher organisms.This study predicts the changes in metabolic flux in both unicellular and multicellular organisms using differences in gene expression and metabolic networks，a method which does not require a priori knowledge or information on the rate of metabolite uptake or secretion of organisms.Methods The present study hypothesized that in a two-state model，significant changes in the expression of an enzyme-encoding gene should result in a significant change in the metabolic rate of the enzyme-catalysed reaction.Microarray analysis of genomic data and whole-genome metabolic networks was used to construct an optimization model that could predict significant changes in the distribution of metabolic flux by maximizing the consistency between changes in metabolic flux and gene expression.Results The model predicted significant changes in metabolic flux，which were consistent with the experimental measurements of metabolic flux changes in aerobic chemostat cultures of Escherichia coli under different dilution rates.Conclusions The present study proposes a constraint-based modelling method that can accurately predict qualitative changes in flux for lower organisms and provides an effective way to investigate the altered metabolism of higher organisms.