Deoiling of petroleum wastewater is a major concern in petroleum process engineering. The conventional design procedure for deoiling systems consists of nine consecutive stages involving different people or groups. This thesis considers reengineering the conventional procedure, with the aim of increasing the efficiency of the process and productivity of people involved. The thesis then investigates the development of the appropriate computer support tools for the new procedure. The reengineering of the procedure is based on concepts proposed by Hammer (1990). The resulting procedure consists of four stages, where parallelism of the different tasks in the conventional procedure is explicitly considered. A model of the new procedure is described in UML (Unified Modelling Language). The advantage of using UML instead of using just natural language description is that UML uses graphical representation that is easy to use and less prone to ambiguity. Based on the new procedure, a decision support system called CODES (Conceptual DEsign of deoiling Systems) was designed and implemented. CODES supports the stages of wastewater assessment, equipment options at different locations, treatment disposal options and recommendations for final process configuration. Two Case Studies using operating field data are reported. The Case Studies using CODES proved to be straightforward to use and produced appropriate comments on the performance of the existing designs and proposed alternatives to satisfy the design requirements. CODES results agreed with knowledge from previous plant trials. The contribution of this thesis to the deoiling design process is twofold. Firstly, a critical review of the conventional procedure has resulted in a new one, which has potential benefits of improving a company's efficiency by reducing the number of people and activities involved, and shortening the design time. Secondly, the feasibility of applying the new procedure and CODES is successfully demonstrated through the case studies. Limitations and areas of extensions are also identified.