Abstract: The measurement of agriculture natural risk has an important theoretical and practical significance for the development of agriculture insurance and the prevention and mitigation of disasters. However, the selection of the method for estimating the risk in the risk assessment is also a basic problem. This paper focus on theoretical analysis of mathematical methods relating to the estimation of agriculture natural (catastrophe) risk, including the “mean-standard deviation” measurement of agriculture losses due to disaster, the fitting method for the distribution of agriculture losses due to disaster and the catastrophe risk measurement model based on the extreme value theory, in a hope of providing scientific basis for the quantitative research on agricultural nature (catastrophic) risk and its insurance pricing．

Key words: agriculture natural (catastrophe) risk, measurement, “mean-standard deviation”, loss distribution, POT model