The authors examined the effects of operating parameters on the $H_2$ production by dark fermentation of the wastewater generated from food waste recycling facilities, in short "food waste wastewater (FWW)". Central composite design based response surface methodology was applied to analyze the effect of initial pH (5.5-8.5) and substrate concentration (2-20 g Carbo. COD/L) on $H_2$ production. The experiment was conducted under mesophilic ($35^{circ}C$) condition and a heat-treated ($90^{circ}C$ for 20min)anaerobic digester sludge was used as a seeding source. Although there was a little difference in carbohydrate removal, $H_2$ yield was largely affected by the experimental conditions, from 0.38 to 1.77 mol $H_2$/mol $hexose_{added}$. By applying regression analysis, $H_2$ yield was well fitted based on the coded value to a second order polynomial equation (p = 0.0243): Y = $1.78-0.17X_1+0.30X_2+0.37X_1X_2-0.29X_1{^2}-0.35X_2{^2}$, where $X_1$, $X_2$, and Y are pH, substrate concentration (g Carbo. COD/L), and hydrogen yield (mol $H_2$/mol $hexose_{added}$), respectively. The 2-D response surface clearly showed a high inter-dependency between initial pH and substrate concentration, and the role of these two factors was to control the pH during fermentation. According to the statistical optimization, the optimum condition of initial pH and substrate concentration were 7.0 and 13.4 g Carbo. COD/L, respectively, under which predicted $H_2$ yield was 1.84 mol $H_2$/mol $hexose_{added}$. Microbial analysis using 16S rRNA PCR-DGGE showed that $Clostridium$ sp. such as $Clostridium$ $perfringens$, $Clostridium$ $sticklandii$, and $Clostridium$ $bifermentans$ were main $H_2$-producers.