In this research, the frequency of frost is analysed from 95 synoptic stations for the period 1990–2015. This information was categorised by a fuzzy c-approach clustering algorithm and indicated that Iran is classified into five clusters with the aid of the frost-occurrence frequencies. The greatest frequency of days with frost prevalence is located in Cluster 1 that consists of Sarab station with an average annual frequency of 141.1 days over the period 1990–2015. The least frequent is found in Cluster 5 that consists of the stations positioned along the south and north coasts. Spatial association for the frequency of incidence of frost days also includes a dependence on elevation and latitude of stations, as well as their situation inside the course of external synoptic systems, bodily and geomorphological features and local climate. Also, a study of daily minimum temperature displays a widespread warming trend at some stage during this period, and has discovered an increase in the index of the number of tropical nights, warmest nights and coldest nights and decreasing trends have been determined in the number of frost days, cool nights and cold spell period index over most regions of Iran.