When designing scenarios for fire safety in residential buildings, one should consider the natural variability of heat release rate of a single item that can possibly influence fire spread between multiple items. This study intends to represent experimentally measured design fire for single residential items using a combination of t-squared growth method and Ingason exponential decay method. The characterizations required three representative data from experiments including maximum heat release rate, time to reach the maximum heat release rate and total heat released. The methods have established good representations of the experimental design fires with more than 86% similarity. Distribution statistics are considered to characterize the design fires in a probabilistic manner which develop possible design fires from an average curve as well as within a minimum/maximum limit and a 5th/95th percentile limit. Comparisons denote that the probabilistic design fires are reliable to represent the possible heat release rate for single residential items.