Multiple water jets and a cross air flow are used to cool a mold segment in a homogeneous and automated manner. The test segment was cooled from an initial temperature of 573 K. An average temperature difference of less than 3 K and a maximum temperature difference of less than 6 K were obtained along the length of the surface of the test segment during the entire duration of the cooling process as opposed to the traditional channel cooling approach where the mean and maximum temperature differences increase over time. The top surface of the test segment represents the mold/part interface which is of interest in this study. Using model predictive control (MPC) and a data-driven predictive model, the cooling speed of the test segment’s top surface was able to be maintained within ±5 K of the cooling ramp imposed. The results were compared to the results obtained when using a simpler On/Off algorithm for automated cooling. Compared to the simpler On/Off algorithm, there was an improvement in the accuracy of the cooling ramp with respect to its reference value of over 30% for most cooling ramps tested (5–25 K/min). A parametric study on the influence of the flowrates of the fluids on the cooling speed of the test segment’s surface was also conducted.