The glass curtain walls have been recently introduced in Tunisia; they are seen as a new fashion and are highly appreciated by some for their pleasing aesthetics. The objective of this paper is to investigate whether the glass curtain walls are appropriate for the Tunisian local climate and context and if it is so, to give recommendations concerning the kind of glass to be used. A TRNSYS [Klein SA. TRNSYS a transient system simulation Program V5 14.2. Solar Energy Laboratory, University of Wisconsin Madison, July 1996] simulation was conducted on a typical administrative building. The investigation concerns only the building heating and cooling load. The building was split in five thermal zones; for each thermal zone, all the windows have the same orientation. The single zone model TYPE19 of TRNSYS [Klein, 1996] was used to model each thermal zone. An additional convection heat transfer between the different thermal zones of the building was modelled according to the Brown and Solvason law [Brown WG, Solvason KR. Natural convection through rectangular openings in partitions, Part 1: vertical partitions. Int. J. Heat Mass Transfer 1962; 5: 859-68]. This particular law was used because it has been validated in the Tunisian context by Bouden [Bouden C. Analyse du suivi thennique d'un pavillon solaire experimental en region Tunisoise, These de doctoral, Universite de Paris 7, 1989]. We assume that the glass curtain wall will be implemented only on the main building facade; this is why it was simulated with different glazing sizes and glass types. The other facades remain unchanged. The results of this simulation have shown that, in relation to space heating, the glass curtain wall can be very interesting in the Tunisian context if the orientation as well as the kind of glazing are carefully selected. (c) 2006 Elsevier Ltd. All rights reserved.