Accumulated chilling was estimated by applying three different models to the hourly autumn-winter temperature records from Santiago (33°34 S lat; 625 m.a.s.l.) and Vicuña (30°02’ S lat; 643 m.a.s.l.) for the years 2005 and 2006. The model of chilling hours, currently used in Chile as an agroclimatic indicator, was of limited use for effectively contrasting a subtropical condition (Vicuña) with a temperate area such as Santiago. The application of the Utah model gave negative values from March to May, and even up to June in Vicuña, since in this model the chilling effect is "negated" by warmer temperatures. However, a modified version of the Utah model named Positive Chilling Units (PCU), in which negative values are omitted, showed differences in the accumulated chilling between both regions, although these differences were of small magnitude and were noted only from July onwards. The Dynamic Model, which considers that chilling is irreversibly accumulated as quantum or Chill Portions (CP), showed that chilling in Santiago doubled that of Vicuña, and that these differences in location were already expressed at the beginning of autumn, confirming, thus, the suitability of the model for subtropical conditions. In this work the advantages of the dynamic model over other models are discussed.