The combination of food quality standard and soil plant transfer models can be used to derive critical limits of heavy metals for agricultural soils. In this paper, a robust methodology is presented, taking the variations of plant species and cultivars and soil properties into account to derive soil thresholds for lead (Pb) applying species sensitivity distribution (SSD). Three species of root vegetables (four cultivars each for radish, carrot, and potato) were selected to investigate their sensitivity differences for accumulating Pb through greenhouse experiment. Empirical soil plant transfer model was developed from carrot New Kuroda grown in twenty-one soils covering a wide variation in physicochemical properties and was used to normalize the bioaccumulation data of non-model cultivars. The relationship was then validated to be reliable and would not cause over-protection using data from field experimental sites and published independent studies. The added hazardous concentration for protecting 95% of the cultivars not exceeding the food quality standard (HC5(add)) were then calculated from the Burr Type III function fitted SSD curves. The derived soil Pb thresholds based on the added risk approach (total soil concentration subtracting the natural background part) were presented as continuous or scenario criteria depending on the combination of soil pH and CEC. (C) 2015 Elsevier B.V. All rights reserved.