Background: The setting of sterilization dose in practice is usually higher than the actual need of product due to the accuracy limit of traditional methods. Extra irradiation dose leads to unnecessary radiation damage to products and reduces the production capacity of irradiation plants. So it is important to find a more accurate method to set the sterilization dose for micropopulation. Purpose: This research is to find a new approach to set the sterilization dose more accurately for micropopulation of products. Methods: First of all

the relationship between the survival bacteria count and radiation dose through theoretical derivation is studied. Then

the data of sterilization dose curve of micropopulation is calculated according to the standard distribution of resistances (SDR). Finally

five kinds of commonly used fitting functions are applied to fit the sterilization dose curve. Results: The five cubed polynomial function is one of the best fitting methods to the sterilization dose curve. The sterilization dose setting method based on five cubed polynomial fitting reduces the deviation range from ?9.8% ? 123.1% to ?7.7% ? 1.9% between set sterilization dose and theoretical value

compared with the traditional method based on linear fitting. Conclusion: The method of polynomial fitting is more accurate that can reduce the extra radiation dose of products effectively when guarantee the quality of sterilization. Abstract