The optimum management of multi-reservoir systems becomes increasingly complex when conflicting uses of water exist, i.e. water supply, irrigation, hydropower generation, etc. This is especially the case in the Nestos basin, a trans-boundary basin extending between Bulgaria and Greece. The transboundary nature of the catchment makes the management of water resources of Nestos complicated. Between 1971 and 1972, the Hellenic Public Power Corporation (PPC) undertook a feasibility study for the construction of three dams. The original plan was to build three multipurpose reservoirs (serially) the first two of which reversible. However the overall project was not completed due to lack of funds, with only two completed reservoirs to this day (Thysavros and Platanovryssi). This paper presents a methodology that combines both simulation and optimization to develop optimal operational rules for the hydrosystem and assess the impact (and potential benefits) from the construction of the third reservoir (Temenos). The main tools used are WEAP21 which simulates the hydrosystem operation and MATLAB where calibration and optimization is taking place. A bespoke code handling the interaction between the two programs was also developed in Matlab, using the COM-API available in WEAP21. Hydrosystem optimization was undertaken using two criteria: the benefit from hydroenergy production and the hydrosystem reliability in terms of water demand coverage. Four scenarios with variation on the operating rules and with or without the Temenos dam were considered and the results compared. The inputs consisted of long synthetic time series to investigate the reliability and responsiveness of the operating rules under uncertainty. Pareto fronts where produced to explore solutions that can achieve the desired reliability in combination with high hydropower production. Based on the work undertaken, the construction of the Temenos reservoir contributes significantly to the increase of hydroenergy production while ensuring that the system will meet water demands. Significant improvement is shown when seasonal variables are used to distinguish between irrigated and non-irrigated periods in the simulation/optimization. Results and operating rules produced appear to be robust under significant long term hydrologic uncertainty with the hydrosystem performing consistently in high reliability levels.