We deal with the problem of verisimilitude , a notion which, roughly speaking, tries to capture how close a scientific theory is to the truth. Our starting philosophical basis is Evandro Agazzi's approach and his view on scientific objectivity which relies on his particular meaning of "partial truth". By following an epistemological approach to the verisimilitude problem and adopting the semantic view of theories, we develop our epistemological proposal about the comparative evaluation of scientific theories and cognitive situations. Our proposal allows to establish, in a qualitative way, in which sense a theory, or a cognitive situation, is better (more verisimilar) than another.