This paper deals with water resource management fuzzy decision systems. The motivation for the study is the need for an automated management policy for an artificial reservoir operating according to common sense rules. As common sense is never defined in terms of strict rules, there is still the possibility to choose the rules so as to follow common sense and to optimize some criteria suitably selected. A hybrid model of the reservoir is considered and implemented in Stateflow/Simulink to test the behavior of the reservoir when the decision strategy is applied. The resulting system is complex enough to motivate the need for efficient nonlinear optimization techniques, hence the parameters of the fuzzy system are optimized by employing genetic algorithms, which have proved very effective in the presence of strong nonlinearity. Moreover, suitable decision variable constraints are considered in order to keep the physical meaning of the rules. Monte Carlo simulations, based on a modified ARMAX model of the inflow, are used to test the effectiveness of the proposed strategy in different operating scenarios. The good results on 10,000 runs show that the proposed strategy not only works on a single realization of the data considered in the study, but it is able to "learn" the characteristics of the considered process.