The influence of thermal stress on health is proved. In the light of the current climate change this relationship should be extensively investigated. This study aims to describe the relationship between temperature and total mortality in the city of Tunis over a period of three years, from 2005 to 2007 using time series analysis. The form of the relation (crude and adjusted) between mortality and temperatures was investigated using Poisson generalized additive models (GAM). Confounders included in the models were pollutant, trend, calendar month, day of the week, the period of Ramadan, and holidays. The adjusted relationship between mortality and temperature was "V" shaped with a steeper slope for low temperatures than for high temperatures. The impact of heat on mortality will be further analyzed to refine our findings and to contribute to the elaboration of guidelines for preventive strategies.