Abstract The linear functions for non-dimensional wind and temperature profiles are commonly used to describe the surface layer fluxes in atmospheric models. However, their applicability is limited to smaller values of the stability parameter z/ L (where z is the height above ground and L is the Obukhov length) i.e. z/ L < 1.0. These linear functions have been modified (Webb 1970, Quart. J. Roy. Meteor. Soc. 96, 67–90; Clarke 1970, Quart. J. Roy. Meteor. Soc. 96, 91–114; Hicks 1976, Quart. J. Roy. Meteor. Soc. 102, 535–551; Beljaars and Holtslag 1991, J. Appl. Meteorol. 30, 327–341; Cheng and Brutsaert 2005, Boundary-Layer Meteorol. 114, 519–538) over the years for calculating fluxes when z/ L > 1.0 under strongly stable conditions. In view of this, the objective of the present study is to analyze the performance of these similarity functions to compute surface fluxes in stable conditions. The meteorological observations from the Cooperative Atmosphere-Surface Exchange Study (CASES-99) experiment are utilized for computing the surface fluxes in stable conditions. The computed fluxes are found to be reasonably close to those observed. The ratio of observed to computed fluxes reveals that the computed fluxes are close to the observations for all the similarity functions for z/ L < 1.0 whereas the computed values show relatively a large scatter from observations for z/ L > 1.0. The computed values of u ⁎ and heat flux do not show significant differences from those observed at 99% confidence limit. The performance of all the similarity functions considered here is found to be comparable to each other in strongly stable conditions.