This research focused on applying predictive models to assess the academic performance of students in the Information Technology (IT) program at the University of the Armed Forces ESPE, Santo Domingo campus, with the aim of identifying the socioeconomic factors that most influence their academic performance, so that the authorities of Ecuadorian higher education institutions with similar characteristics and programs can consider these factors to mitigate the risk of low academic performance. The methodology employed included a logistic regression model to classify the data into categories of good or poor performance, analyzing student attributes such as gender, employment status, family income, among others, using a database that covers the entire student population of the program; a decision tree model, was also incorporated to validate the most influential factors. The innovation in the applied methodology was based on the use of specific characteristics of Ecuadorian IT students, as well as in the validation and comparison of model performance through Principal Component Analysis (PCA) and the application of ANOVA to determine if there are significant differences between the models. The results indicated that non-academic factors, such as pregnancy or illness, along with the father’s occupation, the type of school attended (public or private), and the availability of technological infrastructure, are determinants of the academic performance of IT students; furthermore, it was concluded that both models have high performance and that there are no significant differences between them, or that, if there are any, they are due to chance.