The “predictive modeling for coupled multi-physics systems (PM_CMPS)” methodology is applied in this work to the numerical simulation model of the mechanical draft cooling tower (MDCT) located in the F-area at Savannah River National Laboratory (SRNL) in order to improve the predictions of this model by combining computational information with measurements of outlet air humidity, outlet air and outlet water temperatures. At the outlet of this cooling tower, where measurements of the quantities of interest are available, the PM_CMPS reduces the predicted uncertainties for these quantities to values that are smaller than either the computed or the measured uncertainties. The PM_CMPS has also been applied to reduce the uncertainties for quantities of interest inside the tower’s fill section, where no direct measurements are available. The maximum reductions of uncertainties occur at the locations where direct measurements are available. At other locations, the predicted response uncertainties are reduced by the PM_CMPS methodology to values that are smaller than the modeling uncertainties arising from the imprecisely known model parameters.