The COD removal efficiency from an instant coffee processing wastewater using electrocoagulation was investigated. For this purpose, the response surface methodology was employed, using central composing design to optimize three of the most important operating variables, i.e., electrolysis time, current density and initial pH. The results based upon statistical analysis showed that the quadratic models for COD removal were significant at very low probability value (<0.0001) and high coefficient of determination (R2 = 0.9621) value. The statistical results also indicated that all the three variables and the interaction between initial pH and electrolysis time were significant on COD abatement. The maximum predicted COD removal using the response function reached 93.3% with electrolysis time of 10 min, current density of 108.3 A/m2 and initial pH of 7.0, respectively. The removal efficiency value was agreed well with the experimental value of COD removal (90.4%) under the optimum conditions.