First, using the law of large numbers, we analyze the setting principle of agricultural insurance premium rate, and take the case of setting of adult sow premium rate for study, to draw the conclusion that with the continuous promotion of agricultural insurance, increase in the types of agricultural insurance and increase in the number of the insured, the premium rate should also be adjusted opportunely. Then, on the basis of Bayes' theorem, we adjust and calibrate the claim frequency and the average claim, in order to correctly adjust agricultural insurance premium rate; take the case of forest insurance for premium rate adjustment analysis. In setting and adjustment of agricultural insurance premium rate, in order to make the expected results well close to the real results, it is necessary to apply the probability estimates in a large number of risk units; focus on the establishment of agricultural risk database, to timely adjust agricultural insurance premium rate.