<i>Background:</i> Obtaining accurate height and, consequently, body mass index (BMI) measurements in elderly subjects can be difficult due to changes in posture and loss of height during ageing. Measurements of other body segments can be used as an alternative to estimate standing height, but population- and age-specific equations are necessary. <i>Objective:</i> Our objectives were to validate existing equations, to develop new simple equations to predict height in an elderly Spanish population and to assess the accuracy of the BMI calculated by estimated height from the new equations. <i>Methods:</i> We measured height and demi-span in a representative sample of 592 individuals, 271 men and 321 women, 65 years and older (mean ± SD, 73.8 ± 6.3 years). We suggested equations to predict height from demi-span by multiple regression analyses and performed an agreement analysis between measured and estimated indices. <i>Results:</i> Height estimated from demi-span correlated significantly (p < 0.001) with measured height (men: r = 0.708, women: r = 0.625). The best prediction equations were as follows: men, height (in cm) = 77.821 + (1.132 × demi-span in cm) + (–0.215 × 5-year age category); women: height (in cm) = 88.854 + (0.899 × demi-span in cm) + (–0.692 × 5-year age category). No significant differences between the mean values of estimated and measured heights were found for men (–0.03 ± 4.6 cm) or women (–0.02 ± 4.1 cm). The BMI derived from measured height did not differ significantly from the BMI derived from estimated height either. <i>Conclusion:</i> Predicted height values from equations based on demi-span and age may be acceptable surrogates to derive accurate nutritional indices such as the BMI, particularly in elderly populations, where height may be difficult to measure accurately.