This paper describes in a simple way the development of a web application for predictive maintenance of equipment implemented in the electronics laboratory of the Universidad Pedagógica y Tecnológica de Colombia Sogamoso. This application was developed in Python language to achieve the manipulation and processing of large amounts of data. We developed a machine learning algorithm to predict damages in the laboratory equipment and enable the stakeholder toschedule maintenance to these equipments to prevent them from getting damaged. We implemented and compared the results obtained for two models (Random Forest and MLPRegressor Neural Network), being Random Forest the most accurate model.