Predictive model for estimating total contents of cadmium (Cd) and lead (Pb) was developed by stepwise multiple-regression analysis using chemical properties and extractable contents of metal in paddy soil adjacent to abandoned mines in 2009 and 2010. Soil properties, e.g. pH, electrical conductivity (EC), organic matter (OM), available phosphorus ($P_2O_5$), and exchangeable cations (i.e. Ca, Na, K, Mg) were measured. Total contents of Cd and Pb as well as extractable contents of metals were analyzed by ICP-OES. Results showed that the total and extractable contents were estimated to be 3.55 and $0.27mg;kg^{-1}$ in Cd and 83.38 and $24.32mg;kg^{-1}$ in Pb on the average. From stepwise analysis, it was found that the contents of extractable Cd, Zn, Cu, as well as exchangeable Na were significantly influenced on estimation of the total contents of Cd in soil. Moreover, it also showed that the contents of extractable Pb, Zn, and Cu significantly affected estimation of the total contents of Pb in soil. More significant relationship between estimated and measured value in total contents of Pb was observed than those of Cd ($R^2$=0.87, p<0.0001). This demonstrates that extractable contents of metal are influenced more on estimation of total contents of Cd and Pb in soil than soil properties.