This paper presents an improvement of the CTTC model originally developed by Swaid and Hoffman, which makes it applicable in any fine weather situation. The model accounts for the effect of urban geometry and density on intra-urban variation of air temperatures. The application of the original model as an evaluative tool for practical urban design situations was limited by its restriction in use to fair weather (light wind and clear sky) conditions. In the improved CTTC method most of these restrictions are lifted. The accuracy of the new version is verified against measured data recorded in two canyons located at the University of Adelaide. The predicted temperature proved to be in good agreement with the measured temperature in different synoptic situations (cloudiness and wind speed and direction). An investigation is conducted to test the significance of incorporating the urbanisation effect to modify meteorological air temperature used in thermal simulation programs. The results indicate that it has a crucial effect on predicted energy use for space heating/cooling.