More and more researches are focused on soil quality standards applied for agriculture production of rice in China, but the prediction model based on soil type were the for different for rice cultivars with different accumulation characteristics of toxic elements. This study aimed to develop predication models for cadmium (Cd) transfer from paddy soils across Chinese main rice producing parts to two rice cultivars with extremely high and low ability for accumulating Cd. Thirty-three paddy soils were collected and two rice cultivars were planted with four replicates in the year of 2017 and 2018. “Yelicanghua” is a high grain Cd rice cultivar (YLCH), “Longhuamaohu” is a low grain Cd rice cultivar (LHMH). The results showed that the average brown rice Cd concentrations grown on 33 soils were 0.26 (0.06-0.72) mg kg-1. The average concentrations in rice were 0.14 (0.04-0.55) for “YLCH” and 0.03 (0.01-0.10) for “LHMH” in 2017 and 0.16 (0.02-0.66) for “YLCH” and 0.02 (0.01-0.15) for “LHMH” in 2018, respectively. The predication models were developed using soil pH, organic carbon (OC) and total Cd concentrations. Stepwise multiple linear regression analysis showed that soil Cd and pH were significant variables contributing to high accumulation rice cultivar but soil Cd, pH and OC were significant variables contributing to low Cd accumulation rice cultivar across 33 paddy soils. The predication models were Log[rice-Cd]=-0.097pH+0.602[Total-Cd]+0.078(R2=0.589, n=33, P<0.001) for high Cd cultivar (LHMH) and Log[rice-Cd]=-0.156pH-0.528Log[Total-Cd]+0.445Log[OM] (R2=0.669, n=33, P<0.001) for low Cd cultivars (YLCH), respectively. Then the soil thresholds for the two rice cultivars planting based on its significant variables were deduced from the food safety standard. Deducing soil threshold should take into account ability of accumulating Cd.