A model was developed to classify the Italian territories in relation to their suitability to harbour populations of Culicoides imicola and, as a consequence, also able to sustain a bluetongue (BT) epidemic. Italy was subdivided into 3507 10 × 10 km cells. In 546 cells at least one collection was made. The cell was considered the unit for all subsequent analyses. Culicoides were collected using Onderstepoort-type blacklight traps. Some traps were operated weekly at chosen sites; the remainder were moved almost daily to new sites. Only the results obtained during the peak August–November period were used, to exclude bias caused by the seasonality of C. imicola. Climate data for the period 1999–2001 were obtained from 80 weather stations. Multiple logistic regression was performed using the presence or absence of C. imicola in a specific cell as the dependent variable. Annual means of daily values for minimum temperature and minimum relative humidity, and the mean altitude above sea level, were the independent variables. The probability of occurrence of C. imicola in each grid cell was used to create a prediction map for Italy. The model was able to correctly classify 77.5% of the 546 grid cells in which at least one collection had been made. Culicoides imicola was found frequently through much of Sardinia, in parts of southern Italy, and further north along the Tyrrhenian coast, but was absent from along most of the Adriatic coast, and the internal mainland, and from most of Sicily. Six detailed maps are provided. Also mapped are areas where the probability of the occurrence of C. imicola is lower than 5%. This identification of possible mountainous C. imicola-free areas in central Italy could facilitate safer animal trade and transhumance, even if BT infections in traded animals or moving stock, were to go undetected. Needless to say this depends upon no cool-adapted species of Culicoides being involved in the transmission of BT disease.