The heterologous biosynthesis of 6-deoxyerythronolide B (6dEB), a key intermediate in the biosynthesis of erythromycin, has recently been achieved in Escherichia coli, but the experimental product yield remains low. In this study, in silico strategies were adopted to evaluate and improve the biosynthesis of 6dEB in this strain. The theoretical capability of E. coli to produce 6dEB was first evaluated by analyzing the maximum theoretical molar yield (MTMY) of 6dEB utilizing three carbon sources, glucose, propionate and glycerol. Although propionate is presently most often used experimentally, our results indicated that glucose would be the most feasible substrate for 6dEB production from economic and long-term standpoints. Compared with Saccharomyces cerevisiae and Bacillus subtilis, E. coli was found to be a better heterologous host for the biosynthesis of 6dEB due to the higher MTMY value under the same conditions. Two strategies, including a flux distribution comparison analysis (FDCA) and linear minimization of metabolic adjustment based (LMOMA-based) methods, were proposed and employed for in silico strain improvement of 6dEB production, which yielded several potential gene targets for future experimental validation. In a further analysis, increasing the specific growth rate (SGR) or the non-growth associated maintenance (NGAM) was found to decrease the MTMY; while increasing the specific oxygen uptake rate (SOUR) or the specific carbon source uptake rate (SCUR) increased the MTMY. Taken together, our findings identified key factors directly affecting the MTMY of 6dEB production, which will guide future experimental research or even the industrial production of 6dEB.