Height is an important clinical indicator to derive body mass index (BMI), creatinine height index and also to estimate basal energy expenditure, basal metabolic rate and vital capacity through lung function. However, height measurement in the elderly may impose some difficulties and the reliability is doubtful. Equations estimating height from other anthropometric measures have been developed for Caucasians, but only one study has developed an equation (based on arm span only) for an Asian population. Therefore, a cross sectional study was conducted to develop equations using several anthropometric measurements for estimating stature in Malaysian elderly. A total of 100 adults (aged 30 to 49 y) and 100 elderly subjects (aged 60 to 86 y) from three major ethnic groups of Malays (52%), Chinese (38.5%) and Indians (9.5%) participated in this study. Anthropometric measurements included body weight, height, arm span, half arm span, demi span and knee height were carried out by trained nutritionists. Inter and intra observer errors and also % Coefficient Variation (%CV) were calculated for each anthropometric measurement. Equations to estimate stature were developed from the anthropometric measurements of arm span, demi span and knee height of adults using linear regression analysis according to sex. Elderly subjects were shorter and lighter compared to their younger counterparts. The %CV of anthropometric measurements in adults and elderly subjects ranged between 5 to 6%, with standing height having the lowest %CV. When the equations derived from adults were applied to elderly subjects, it was found that percentage difference between actual height and the estimated value ranged from 1.0 to 3.3%. However, the percentage difference between estimated height from the equations developed in this study compared to those derived from the equations of other populations ranged between 0.2 to 8.7%. In conclusion, standing height is an ideal technique for estimating the stature of individuals. However, in cases where its measurement is not possible or reliable, such as in elderly subjects, height can be estimated from proxy indicators of stature. In this study arm span showed the highest correlation with standing height, which is in agreement with other studies. It should be borne in mind that equations derived from taller statured populations (e.g. Caucasians) may be less accurate when applied to shorter statured populations.