A novel synthesis approach is proposed for designing integrated processes for oily wastewater treatment. A mixed-integer linear programming (MILP) model was developed based on the superstructure of alternative separation processes. These processes are grouped into three main sets, i.e. the pre-treatment, intermediate treatment, and final treatment, from which the optimal combination of technologies is selected in order to achieve the legally required effluent quality. The parametric analysis was performed by varying the input chemical oxygen demand (COD) values of oil-in-water emulsion between 1000 and 110 000 mg/L, yielding a variety of combinations of purification techniques. The optimal designs obtained indicate that oil-in-water emulsions can be treated successfully and profitably by integrated processes composed of the established separation techniques.