Drinking water treatment plants (DWTPs) have to be efficiently managed to produce safe water at all times, independently from variations of influent water quality. Artificial Intelligence techniques like fuzzy inference systems (FIS) can help in consolidating process knowledge accumulated through years of experience and improve the consistency and resiliency of decision-making in these facilities. The objective of this study was to develop an advanced control system for choosing the combined dose of sodium hypochlorite and chlorine dioxide at the primary disinfection step of a full-scale DWTP. To accomplish this, two FISs consisting of a feed-forward and feed-back control elements were developed. The models were integrated in an environmental decision support system (EDSS) that evaluates the disinfection by-products (DBPs) formation risk and proposes actions which can be verified and applied by treatment managers. Implementation of the EDSS at a full-scale DWTP of Barcelona was positively validated 85.5 % of the times, maintaining acceptable DBPs concentrations at the effluent. The presented methodology can be used at similar surface water DWTPs for developing control strategies to manage DBPs formation.