The objective of this paper is to compare the prediction capability of different correlations for calculating delayed coking yields. The evaluation was developed taking operation data reported in the literature into account for delayed coking commercial plants. The effects of pressure, feed type, and temperature on product yields were analyzed. Correlations that include the effect of operating conditions proved to be more accurate compared to those that consider only feed properties. From the calculation of yields, it is possible to condude that an increase of 1 wt % of coke yield is obtained for each 1 wt % increase of the feed Conradson carbon residue (CCR) or for each 5 psig increase of the coke drum pressure and a reduction of 1 wt % of coke yield is achieved for each 15 F increase of the coke drum temperature. The correlation developed by Volk et al. resulted to be the most accurate correlation to predict coke yields, while the most popular correlations (Gary-Handwerk and Maples) are the worst.