A set of experiments was carried out to maximize adenosine 5'-triphosphate (ATP) extraction efficiency from acidogenic culture using whey wastewater. ATP concentrations at different microbial concentrations increased linearly as microbial concentration decreased. More than 50% of ATP was extracted from the sample of 39 mg volatile suspended solids (VSS)/l compared to the sample of 2.8 g VSS/l. The ATP concentrations of the corresponding samples were 0.74+/-0.06 and 0.49+/-0.05 mg/l, respectively. For low VSS concentrations ranging from 39 to 92 mg/l, the extracted ATP concentration did not vary significantly at 0.73+/-0.01 mg ATP/l. Response surface methodology with a central composite in cube design for the experiments was used to locate the optimum for maximal ATP extraction with respect to boiling and bead beating treatments. The overall designed intervals were from 0 to 15 min and from 0 to 3 min for boiling and bead beating, respectively. The extracted ATP concentration ranged from 0.01 to 0.74 mg/l within the design boundary. The following is a partial cubic model where eta is the concentration of ATP and x ( k ) is the corresponding variable term (k=boiling time and bead beating time in order): eta=0.629+0.035x (1)-0.818x (2)-0.002x (1) x (2)-0.003x (1) (2) +0.254x (2) (2) +0.002x (1) (2) x (2). This model successfully approximates the response of ATP concentration with respect to the boiling- and bead beating-time. The condition for maximal ATP extraction was 5.6 min boiling without bead beating. The maximal ATP concentration using the model was 0.74 mg/l, which was identical to the experimental value at optimum condition for ATP extraction.