Risk prediction scores have received much attention the past few years, especially in the primary prevention of cardiovascular disease (CVD). Although diet has been independently associated with CVD risk, its role in the accuracy of the developed scores has rarely been studied. Thus, in this review, the role of diet assessment on the performance of CVD risk scores and models was critically discussed. A computer-assisted literature search retrieved 15 relevant studies, but only two out of them evaluated the role of diet on the accuracy of the developed models; the inclusion of diet assessment improved significantly the accuracy of CVD risk models. The remaining studies suggested an independent, protective effect of healthy dietary habits on CVD risk, with an attributable risk varying from 9 to 37%. Inclusion of diet component in CVD risks scores, may increase the accuracy of the models, and better identify people at high risk.