This paper presents a novel approach to temperature probability density distribution and function. Probability density functions and frequency are successfully used in wind speed and solar energy analyses in literature. This study applies these data to temperature data analysis. The present model is developed using the indoor and outdoor temperature as a parameter. Outdoor temperature distribution is crucial for the calculation of monthly and total degree-hour. In this paper, using past weather data, the outdoor temperature probability density functions are modeled for four cities in different regions in Turkey via a new computer program. The main advantage of this approach is to allow us to determine heating and cooling loads with respect to different indoor and outdoor temperatures. (C) 2010 Elsevier Ltd. All rights reserved.