In order to predict the consequences and value of frost adaptation through breeding and agronomy across Australia’s cropping region it is essential that a validated frost damage function is incorporated into our crop models. This paper reports on the development of an empirical relationship that predicts the frost induced sterility in wheat at anthesis developed from the analysis of GRDC National Frost Initiative (NFI) data collated from three Australian States. The analysis showed that the extent of frost damage to individual heads was primarily a function of the degree of coldness and its duration. A single function for frost damage to wheat at flowering was derived for all cultivars. While individual frost damage functions were created for each cultivar, with the available data the increased complexity did not appear to improve the damage function. In part this was due to the fact that individual heads were not subjected to the same frost events, with only a few larger frost events for each individual cultivar. This study was a good start in improving the conceptualisation and empirical relationships which could be used in frost models. However, further development is needed to scale the empirical relationship of damage across multiple stages of crop development and from an individual head to a paddock of wheat where plants and individual heads vary in their phenology.