The incomplete knowledge we have regarding the mechanisms involved in the pain experience, especially if it expresses in its pathological form, demands an interdisciplinary approach that includes the use of mathematical and computational models. Some advantages of this approach lies in its non-invasive character and its ability to allow both the demonstration of hypotheses and theories, and the development of new ones. However, the use of computational models for the study of pain (CMSP) seems to be very limited in comparison with that given to more traditional methods, which does not correspond to the holistic approach that is required to address the multidimensional nature of pain. In this paper, we review and discuss the different strategies that have been used in the construction of CMSP, in order to identify some tendencies that possibly have prevented a greater use of them in clinical practice and to propose, eventually, certain recommendations that may be useful to address those issues.