Academic fees are an annual amount that students must pay in return for the training they receive. These fees can change according to several factors. Among these factors, some can influence the increase of tuition fees, while others can ensure that fees are moderate. Therefore, this research aimed to analyze these factors and integrate them as predictors in a deep-learning model for predicting university tuition fees. Hence, the authors used quantitative analysis based on secondary data on tuition fees at the Université de l’Assomption au Congo (UAC). These data were used to develop two regressive neural network models, namely the three-hidden-layer neural network and the four-hidden-layer neural network, to determine the best model for prediction and deployment purposes. The metrics used to evaluate the performance of these two models were mean absolute error, mean square error, root mean square error and coefficient of determination. The results revealed that academic costs increase as a student moves up the promotion ladder. After developing these models, although they all performed successively at 95.3% and 95.6%, the hidden 4-layer model was deployed. The predictors used as features were six: academic year, promotion, tuition fees, dissertation fees, partial time lecturers fees and equipment fees.