Sustainable management of drinking water reservoirs requires balancing the demands of water supply whilst minimizing environmental impact. This study numerically simulates the effect of an improved withdrawal scheme designed to alleviate the temperature pollution downstream of a reservoir. The aim was to identify an optimal withdrawal strategy such that water of a desirable discharge temperature can be supplied downstream without leading to unacceptably low oxygen concentrations within the reservoir. First, we calibrated a one-dimensional numerical model for hydrodynamics and oxygen dynamics (GLM-AED2), verifying that the model reproduced water temperatures and hypolimnetic dissolved oxygen concentrations accurately over a 5 year period. Second, the model was extended to include an adaptive withdrawal functionality, allowing for a prescribed withdrawal temperature to be found, with the potential constraint of hypolimnetic oxygen concentration. Scenario simulations on epi-/metalimnetic withdrawal demonstrate that the model is able to autonomously determine the best withdrawal height depending on the thermal structure and the hypolimnetic oxygen concentration thereby optimizing the ability to supply a desirable discharge temperature to the downstream river during summer. This new withdrawal strategy also increased the hypolimnetic raw water volume to be used for drinking water supply, but reduced the dissolved oxygen concentrations in the deep and cold water layers (hypolimnion). Implications of the results for reservoir management are discussed and the numerical model is provided for operators as a simple and efficient tool for optimizing the withdrawal strategy within different reservoir contexts.