The main purpose of formulating methodologies for building systems' evaluation in low-cost housing is to find an effective solution for the huge Brazilian housing deficit of approximately five million housing units, mainly due to an accelerated population growth in urban centers. Low-cost housing programs are usually implemented in a broad sense, with no regard to local specific conditions. Thus, building systems of quite similar characteristics are employed in places with different climatic conditions, which leads to low-quality houses that do not respond to the users' needs. In this paper, the results of the application of formulas to predict daily indoor temperatures in three monitored low-cost houses in Curitiba, Brazil, are presented. The houses were occupied by families having neither cooling nor heating devices and are built of different building materials with different thermal properties. The monitoring of the houses took place both in winter and in summer. Measured data were also compared with simulated data. In this case, the French software COMFIE was used. Finally, the results of the thermal simulations were compared with those of predictive formulas developed by Givoni. (C) 2004 Elsevier B.V. All rights reserved.