To improve the quality of dried rice seed and solve the application limitation in existing models for predicting germinating rate of dried paddy, we conducted experiments of paddy drying in a deep fixed-bed by using a quadratic orthogonal rotating design. First order dynamic model for predicting germinating rate of paddy dried in deep fixed-bed was developed and verified, model parameters were optimized: undetermined constants of Z1=0.06011 and Z2=0.02558, activation energy of germ protein denaturation E=194.18333. Model examination showed that higher prediction accuracy was achieved with error square sum 0.48358. A regression model describing relations between experiment factor and germinating rate was developed, and impact of experiment factors on germinating rate was analyzed. The order of the factor was clarified though calculation of contribution rate. The two types of model were compared and the parameters for paddy drying were optimized by frequency analytical method. An optimum range of the parameters for germinating rate higher than 90% was given with 95% probability. The results provide useful references for parameter design of drying equipment and drying procedure of rice seed in practice.