At Higher Educational Institutions (HEI) the high hierarchical managers and directors face many challenges during the decision-making process, that sometimes are rely on intuition, and past experiences, leading not just to delays but the low impact in the whole academic community. A decision-making model for managers and administrator of HEIs is presented. We propose a detailed methodology when academic prognosis is taking place. The comparison between five robust Machine Learning algorithms is executed accomplishing outperformed results by Support Vector Machine. As a validation experiment, we executed the proposed decision model in a face-to-face public university in Colombia, showing the results in a developed web platform prototype with its correspondent architecture. Moreover, we discuss the social implication of low graduation rates.