Defined medium for arachidonic acid (ARA) production by Mortierella alpina was optimized for its metabolomics study. For this purpose, a visualization method (VM) was applied for the first time. Experiments were designed according to the uniform design with four factors (concentrations of glucose, NaNO3, KH2PO4 and MgSO4 center dot 7H(2)O) for each at nine levels. Dry cell weight (DCW), ARA yield in DCW [percent (w/w)] and ARA content in total fatty acids [percent (w/w)] were considered as the three objectives. Optimization of single-objective function and multi-objective function of two objectives and three objectives was attempted. Optimal DCW, ARA yield and ARA content were predicted to occur in a medium that contained (grams per litre): glucose 35, NaNO3 1, KH2PO4 7.5 and MgSO4 center dot 7H(2)O 2.6. Upon verification, the average tested DCW (12.95 g/l), ARA yield (18.89 %) and ARA content (42.36 %) were fairly close to the predicted values (12.88 g/l, 9.68 % and 35.57 %, respectively). Moreover, DCW, ARA yield and ARA content from the optimum medium increased by 35.68, 47.23 and 30.90 % compared with control, respectively, indicating that VM had succeeded in exploiting the biomass growth and ARA production by M. alpina.