In this work, a control strategy to guarantee a minimum dissolved oxygen concentration in aerobic fed-batch bioprocesses is proposed. The strategy is equally applicable to avoid low oxygen concentrations that might be detrimental for metabolites production, to prevent metabolic changes and to maintain an oxygen level in the culture compatible with the desired growth rate during the growing phase of the process. It essentially consists in adapting the set-point of the feeding controller to cope with faults and saturation in the oxygen regulation subsystem. The idea behind this adaptation is to manipulate, when necessary, the oxygen consumption in such a way that the oxygen in the culture is maintained above a prescribed threshold. The algorithm is based on sliding mode techniques, and its main features are as follows: simplicity, robustness to process uncertainties, easy implementation, and versatility to be combined with different dissolved oxygen and feeding controllers. The method is assessed numerically in a fed-batch growth phase of Pichia pastoris where an oxygen shortage leads to excessive glycerol accumulation.