The purpose of this study is to find alternatives to conventional temperature-regulation systems using machine learning techniques. This is done by training several machine learning algorithms to regulate the temperature in a simulated factory, the models are then evaluated by making them regulate in said simulated factory. The study found that neural network regressors are best suited to the task. The implications of the study are both practical and academic. Academically, the study could be used as a basis for further study. Practically, it can be used to implement solutions based on its findings.