The commonly used predictive equations to calculate resting energy expenditure (REE) can yield inaccurate results. Aim: To compare measured REE, with estimated REE in normal and obese adults. To develop specific predictive equations for our population. Patients and Methods: Eight hundred sixteen women aged 18 to 74 years old with a body mass index (BMI) between 18.5 and 69.7 kg/m² and 441 men aged 18 to 71 years old with a BMI between 185 and 67.9 kg/m2 were studied. REE was measured by indirect calorimetry and fat free mass by tetrapolar bioimpedance. REE was also estimated using FAO/WHO (F/W) and Harris-Benedict (H/B) equations. Results: Measured REE in controls was 20.7 ± 1.9 and 23.6 ± 3.3 kcal/kg/day in women and men respectively. The figures for overweight women and men were 19.8 ± 1.9 and 20.0 ± 2 kcal/kg/day, for obese women and men were 18.3 ± 1.7 and 19.0 ± 1.8 kcal/kg/day and for morbidly obese women and men, were 16.9 ± 1.8 and 16.1 ± 1.7 kcal/kg/day. When REE was corrected for fat free mass, no differences between controls and different degrees of obesity, were observed. The difference between estimated and measured REE ranged from -420 to 617 kcal in women and from -400 to 900 kcal in men. The equations derived using data obtained in the present study, showed a better predictive accuracy. Conclusions: An important error was detected when REE was predicted using FAO/WHO or Harris Benedict equations, in Chilean obese subjects. Therefore these equations must be used with caution, local equations must be devised or resting energy expenditure must be measured by indirect calorimetry (Rev Méd Chile 2002; 130: 51-60)