Sentiment analysis studies people’s feelings and attitudes present in natural language texts. In education, the application of sentiment analysis has proven to be relevant in understanding students’ perceptions of various aspects of the learning process. Although numerous studies have examined the applicability of sentiment analysis in higher education, the focus has predominantly been on English and Chinese texts. As in other domains, there is a clear lack of state-of-the-art models fine-tuned for sentiment analysis tasks on Spanish texts in higher education. In this study, we propose a framework to address this lack of models, considering their potential utility in other domains with analogous conditions. To this end we chose and evaluated three sentiment analysis tools, that use models either pre-trained in Spanish or fine-tuned for higher education (HE) in other languages, using three domain-specific datasets. The results are promising, with one tool achieving performance comparable to prior studies in English and Chinese. However, challenges emerged during the evaluation, prompting a discussion on these issues and recommendations for future research improvements.