Biofuel production from microalgae requires optimizing the operation of cultivation systems (i.e. outdoor ponds) for this process to be economically sustainable. Controlling algal ponds is complex as the cultivation system is exposed to fluctuating conditions. The strategy investigated in this study uses weather forecast coupled to a predictive model of algal productivity to optimize pond operation. The selected controlled variables were the rates of fresh medium injection and culture removal into and from the pond. This optimization strategy was applied at two locations in France and was shown to increase the productivity by a factor 1.7-2.4 compared to where the pond depth and dilution rate were kept constant over time. A thorough analysis of the optimizer behavior showed that this increase of productivity was achieved by ‘flushing’ the pond and controlling the pond depth. These mechanisms allowed maintaining the biomass concentration and the pond temperature near their optimal values. The complex behavior of the optimizer was reduced to six simple rules that could be used as guidelines for practical operation.