Steam generators are complex equipment and a proper operation depends on the identification of their mostrelevant parameters. The present paper presents an approach to select and rank the operational parameters of a subcrit-ical steam generator of an actual 360 MW Brazilian power plant fueled by Colombian coal. The steam generator andits three coal mills are gathered on a single system whose controllable parameters are identified. Statistical tools likeDesign of Experiments (DoE) and Response Surface Methodology (RSM) are used to identify the model main controllableparameters and interactions to then rank them by order of importance. The surrogate model based on the commercialsoftware is built to simulate the system efficiency with seven controllable input parameters: primary air flow, pulverizedcoal outlet temperature, speed of the dynamic classifier, stoichiometry, excess O2, secondary and primary air crossoverduct pressure, ranked by descendent significance. The maximum relative deviation of that surrogate model compared tothe software simulation is 0.0172. The operator attention must be kept on the most influential parameters, respectingtheir rank and initializing the alterations for a new condition always for the controllable parameters with a high effecton the steam generator efficiency. Finally, controllable parameters must attain the best operating ranges propose. Therecommended operational ranges and order of operation by significance allows a precision action in order to achieve anew, safe, stable, and more efficient condition.