<p>Composite materials such as polymer-reinforced carbon fiber have been increasingly used in various sectors due to their reduced mass and high mechanical strength. The aeronautical sector, for example, has shown that manufacturing the Boeing 787 with 50% of its structure using this type of material led to fuel savings of 25%. However, when using these materials, machining is used as a secondary manufacturing process for geometry adjustments and can represent considerable costs for the manufacturing of components. Developing tools that can assist in real-time control of the surface quality of these machined parts is essential to understand, control, and optimize the machining process. Few studies have been found successfully associating some machine learning models with surface roughness, and the vast majority are focused on metalsand using their own machining parameters. This study investigates whether considering also the acoustic emission signal emitted during the machining process would help or not to improve the surface roughness prediction. For that, this study applies a set of more than twenty machine learning models tuned by Bayesian optimization to a dataset constructed using optimal designof experimentsfor the milling of polymer-reinforced carbon fiber under fourmachining variables (cutting speed, tool condition, milling direction, and carbon fiber) and fifteen input parameters from acoustic emission analysis. Our studies show that it is possible to successfully predict surface roughness with these parameters, showing the best machine learning algorithm and its hyperparameters for this purpose, and the six most relevant features out of nineteen.&#xa0;</p>