Height measurement is an important part of nutritional assessment especially in children. However, in such cases as hospitalized or certain kinds of malformations or disabilities, height cannot be measured accurately. We aimed to determine appropriate height predictors in Iranian healthy children for further use in disabled and/or hospitalized children. A total of 730 apparently healthy children aged 7-11 years old from both sexes from Tehran, Meshed and Rasht were enrolled in a cross sectional study. Height, demispan (DS), halfspan (HS), arm length (AL) and tibia length (TL) were all measured using a measuring tape. Linear regression models were established between height, DS, HS, AL and TL. For boys AL (R2=0.783) and TL (R2=0.837) and for girls AL (R2=0.720), TL (R2=0.765), HS (R2=0.771) and age (R2=0.775), respectively, entered the linear regression model. When height predictors were evaluated individually for each city, only in Tehran DS also entered the regression model. Concordance of different percentiles of height estimates based on AL with those of actual height proposed this measure as a reliable height proxy for this age group in clinical as well as field practice.