An influence diagram (ID) is a method of graphical representation of uncertain knowledge, which can be employed to support decisions in health care using probabilistic reasoning. We aimed to describe the development of an ID to support the decision-making process in phase II at Cardiopulmonary and Metabolic Rehabilitation Program (CPMR). The development of the ID was carried out through the identification of relevant variables and their possible values, as well as the identification of details of each variable, in order to find a network structure that appropriately connects the nodes that represent the variables, with arcs linking acyclic graphs, and to build the graph using specialized knowledge and the conditional probability table for each node in the graph. In spite of the complexity of the interactions, the model obtained with the ID seems to contribute in the decision-making process in phase II CPMR, providing a second opinion to the health pratictioner and helping in diagnostic, therapeutic and decision-making processes, since it is useful in situations with non-linear modeling or with absent or uncertain information.