Sequential experimental design strategy was applied in order to investigate the best conditions for tiger nut milk extraction process taking into account performance and physicochemical features, simultaneously. Protein, lipid, starch, calcium and total sugar composition were considered for this optimization. The tested factors were: sodium metabisulphite concentration (X 1 ), soaking time (X 2 ), soaking temperature (X 3 ) and water proportion (X 4 ). From the 2 4-1 fractional factorial design all independent variables were found to be statistically significant (p <0.1) and X 1 , X 2 and X 3 were chosen for the central composite design as water proportion was already in the minimum value. The best operational conditions, obtained by desirability function, were: 8 h of soaking time, 22° C for soaking temperature, 0.20% (w/v) of sodium metabisulphite and 1:1 tuber (g): water (ml) proportion. In those conditions a 78% yield was obtained for a tiger nut milk with 2.60% protein, 3.7% total sugar and 8.72% starch. The high yield obtained and the great nutritional value shows that this milk has a great potential for application in the food industry.