Moral Foundations Theory is a socio-cognitive psychological theory that constitutes a general framework aimed at explaining the origin and evolution of human moral reasoning. Due to its dyadic structure of values and their violations, it can be used as a theoretical background for discerning moral values from natural language text as it captures a user’s perspective on a specific topic. In this paper, we focus on the automatic detection of moral content in sentences or short paragraphs by means of machine learning techniques. We leverage on a corpus of tweets previously labeled as containing values or violations, according to the Moral Foundations Theory. We double evaluate the result of our work: (i) we compare the results of our model with the state of the art and (ii) we assess the proposed model in detecting the moral values with their polarity. The final outcome shows both an overall improvement in detecting moral content compared to the state of the art and adequate performances in detecting moral values with their sentiment polarity.