A casual or late diagnosis given the lack of knowledge of the severity of a problem or asymptomatology of patients with comorbidity can cause serious consequences for the patient. Therefore, there is an urgent need for decision-making systems based on continuous monitoring in real time. This research presents SmartNextGISSA, an intelligent mechanism used in NextGISSA, an architecture developed with the objective of assisting decision making in a situation room in monitoring patients with comorbidity, based on the crossing of data collected in real time from this patient and his health history. NextGISSA predictively and continuously monitors in real time patients with comorbidities (Pregnant Women, Hypertensives, Diabetics and Bedridden) of GISSA, an intelligent governance platform for decision-making in Digital Health, a product already operational in several municipalities by the startup “Avicenna Intelligent Governance”. With the intention of validating the machine learning module of the architecture proposed in NextGISSA, the SmartNextGISSA module was implemented, as an emphasis of this article, which is characterized by a composition of the NextGISSA architecture responsible for the inference about the data of an individual classifying it with different comorbidity risk levels through machine learning algorithms. Thus, we present a development pipeline that consist of a comparative analysis between binary classification algorithms, such as Naive Bayes, Decision Tree and Random Forest, using the CRISPDM methodology, a concept applied to data mining processes and, later, the model was implemented in a web application.