With Industry 4.0, the integration between physical and digitalenvironments enabled some improvements within several productionsegments. Among these improvements, advancements on theapplication of machine learning algorithms to predict current andfuture states of equipment have been gaining attention, specially,for maintenance purposes. This research work presents a comparativeexperimental study on machine learning algorithms appliedto classification of industrial machinery states. After training andevaluating models based on five different algorithms (i.e., DecisionTree, Naive Bayes, Support Vector Machines, XGBoost and NeuralNetwork), some interesting results were obtained. Considering theaccuracy, precision, recall and training time of each model, it wasobserved that some models performed well, while others may notbe as suitable for solving the problem. Such good performing modelscould be used to schedule interventions on a given industrialequipment, avoiding production stoppages.