Abstract: Estimating accurate insurance rate is very important to promote the development of crop insurance, it can contribute to achieving the goal that farmers can afford the crop insurance, insurance companies can afford the compensation and government can afford the subsidy. First, we used Normal distribution, Chi-square distribution, Weibull distribution, Logistic distribution, Gamma distribution, Extreme (Min) distribution and nonparametric kernel density distribution to simulate the probability distribution of the loss rate of rice yield in Fujian, and estimated its pure rates accordingly. Second, we used experience rating method to calculate the rice pure rate. Finally, based on the eight pure rates, we used the method of arithmetic average to estimate the rice yield insurance pure rate of Fujian, the pure rate was 1.2922%, and according to this rate, calculated the rice insurance rates. The results will provide reference for setting rice area insurance rate of Fujian.