Objective Equations are frequently used to predict resting energy expenditure (REE) in clinical settings. This study aimed to evaluate the accuracy of the eight prediction equations for REE by using Bland-Altman method in a population of end-stage renal disease (ESRD) patients on maintenance hemodialysis (MHD). Methods In a total of 53 ESRD patients on MHD, REE was measured by indirect calorimetry, and then compared with the values calculated from the eight prediction equations suggested by the Food and Agriculture Organization/World Health Organization/United Nations University (FAO/WHO/UNU), Schofield, Harris-Benedict, Owen, Mifflin, Cunningham, Liu, and Hong Jia. Agreement between the REE from indirect calorimetry and the values calculated from the prediction equations was assessed by Spearman correlation, paired t-test, Bland-Altman method and the percentage of individual coincidence. Results The measured REE was 1460±398kcal/day. There were significant discrepancies between the measured REE and the values predicted from Mifflin, Liu, and Hong Jia equations (P<0.05), among which that predicted from Hong Jia equation (240±321kcal/day) was found to have the maximal deviation from the measured REE. There were no significant discrepancies between the measured REE and the values predicted from FAO/WHO/UNU, Schofield, Harris-Benedict, Owen and Cunningham equations (P>0.05), among which the value from Schofield equation (6±293kcal/day) showed the minimal deviation from the measured REE. When Bland-Altman method was used to calculate the 95% limits of agreement, however, the agreement between the measured REE and the values predicted from the 8 equations was very weak. The REE predicted from Schofield equation showed the smallest bias (the low limit was -580±137kcal/day, and the high limit was 592±137kcal/day) and the greatest percentage of individual coincidence (41.5%). Conclusion No equation can be used to accurately predict REE. We therefore recommend indirect calorimetry as the method for determining REE in MHD patients. Schofield equation appears to be the most reliable one among the eight equations. We can use this equation when indirect calorimetry is unavailable.